Issuu on Google+

Praise from the Experts "The author is to be congratulated and appreciated for writing such a practical and useful book for both entry-level and seasoned practitioners. It provides very clear step-by-step instructions using various case examples. In addition, it also provides the explanation of essential theoretical background in each modeling technique." Carl Lee, Ph.D. Professor of Statistics Data Mining Program Coordinator Central Michigan University

"Kattamuri Sarma has written a practical guide that will be o f use to modeling professionals of all backgrounds and levels of experience. In it he shows us the power o f SAS Enterprise Miner to enhance the modeling process in a way that is straightforward, easy to follow, and thorough. From data scrubbing to model fitting, each step in the model building process is described in a clear manner. The book draws on the author's extensive real-world experience to illustrate the ways in which SAS Enterprise Miner can help to facilitate the modeling process." Lee Medoff SAS User 'The content is wonderful, clear, and thorough.' Andrea Wainwright-Zimmerman Senior Statistical Analysis Manager

" I think this book will be very helpful to new users o f SAS Enterprise Miner. It presents many examples in a way that puts the material across clearly. I especially liked the chapters on decision trees and neural networks for their detailed exposition. I also was pleased to see many cases where the author shows the SAS code behind the scenes. I think Dr. Sarma did a fine job." Rich Perline, Ph.D. Senior Scientist NuTech Solutions, Inc.


Predictive Modeling with SAS* Enterprise Miner™ Practical Solutions for Business Applications

Kattamuri S. Sarma, Ph.D.


The correct bibliographic citation for this manual is as follows: Sarma, Kattamuri S. 2007. Predictive Modeling with SAS* Enterprise Miner": Practical Solutions for Business Applications. Cary, NC: SAS Institute Inc.

Predictive Modeling with SAS* Enterprise Miner ": Practical Solutions for Business Applications Copyright Š 2007, SAS InstituteInc., Cary, NC, USA ISBN 978-1-59047-703-8 All rights reserved. Produced in the United States of America. For a hard-copy book: No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, or otherwise, without the prior written permission of the publisher, SAS Institute Inc. For a Web download or e-book: Your use of this publication shall be governed by the terms established by the vendor at the time you acquire this publication. U.S. Government Restricted Rights Notice: Use, duplication, or disclosure of this software and related documentation by the U.S. government is subject to the Agreement with SAS Institute and the restrictions set forth in FAR 52.227-19, Commercial Computer Software-Restricted Rights (June 1987). SAS Institute Inc., SAS Campus Drive, Cary, North Carolina 27513. 1st printing, September 2007 2nd printing, May 2009 3rd printing, December 2009 SAS* Publishing provides a complete selection of books and electronic products to help customers use SAS software to its fullest potential. For more information about our e-books, e-learning products, CDs, and hard-copy books, visit the SAS Publishing Web site at support.sas.com/publishing or call 1-800-727-3228. SAS* and all other SAS Institute Inc. product or service names are registered trademarks or trademarks of SAS Institute Inc. in the USA and other countries. Ž indicatesUSA registration. Other brand and product names are registered trademarks or trademarks of their respective companies.


Contents Preface Acknowledgments

Chapter 1 Research Strategy 1.1 1.2 1.3 1.4 1.5 1.6

Introduction Measurement Scales for Variables Defining the Target Sources of Modeling Data Pre-Processing the Data Alternative Modeling Strategies

Chapter 2 Getting Started w i t h Predictive Modeling 2.1 Introduction 2.2 Opening SAS Enterprise Miner 5.2 2.3 Creating a New Project in SAS Enterprise Miner 5.2 2.4 The SAS Enterprise Miner Window 2.5 Creating a SAS Data Source 2.6 Creating a Process Flow Diagram 2.7 Summary 2.8 Appendix to Chapter 2

vii xi

1 1 2 2 11 12 14

17 17 18 20 21 22 35 72 72

Chapter 3 Variable Selection and Transformation off Variables 77 3.1 3.2 3.3 3.4 3.5

Introduction Variable Selection Transformation of Variables Summary Appendix to Chapter 3

Chapter 4 Building Decision Tree Models t o Predict Response and Risk

77 78 99 109 110

113

4.1 Introduction 113 4.2 An Overview of the Tree Methodology in SAS Enterprise Miner 114 4.3 Development of the Tree in SAS Enterprise Miner... 121


iv Contents

4.4 A Decision Tree Model to Predict Response to Direct Marketing 4.5 Developing a Regression Tree Model to Predict Risk 4.6 Summary 4.7 Appendix to Chapter 4

Chapter 5

Neural Network Models t o Predict Response and Risk 5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8 5.9

143 159 165 166

169

Introduction A General Example of a Neural Network Model Estimation of Weights in a Neural Network Model... A Neural Network Model to Predict Response A Neural Network Model to Predict Loss Frequency in Auto Insurance An Introduction to Radial Basis Functions Alternative Specifications of the Neural Network Architecture Summary Appendix to Chapter 5

Chapter 6 Regression Models 6.1 Introduction 6.2 What Types of Models Can Be Developed Using the Regression Node? 6.3 An Overview of Some Properties of the Regression Node 6.4 Business Applications 6.5 Appendix to Chapter 6

Chapter 7 Comparison of Different Models 7.1 Introduction 7.2 Models for Binary Targets: An Example of Predicting Attrition 7.3 Models for Ordinal Targets: An Example of Predicting Accident Risk 7.4 Comparison of All Three Accident Risk Models

169 172 180 181 204 216 219 231 232

235 235 236 249 273 301

305 305 305 316 327


Contents v

Chapter 8 Customer Profitability 8.1 8.2 8.3 8.4 8.5 8.6 8.7 8.8 8.9

Glossary..... References Index

329

Introduction Acquisition Cost Cost of Default Revenue Profit The Optimum Cut-off Point Alternative Scenarios of Response and Risk Customer Lifetime Value Suggestions for Extending Results

-

329 331 333 334 334 336 337 338 338

339 347 349


Preface In order to make effective use of the tools provided in SAS Enterprise Miner, you need to be able to do more than just set the software in motion. While it is essential to know the mechanics o f how to use each tool (or node, as Enterprise Miner tools are called), you should also understand the methodology behind each tool and be able to interpret the output produced by each tool and the multitude of options available with each one. Although this book will appeal to beginners because of its step-by-step, screen-by-screen introduction to the basic tasks that can be accomplished with Enterprise Miner, it also provides the depth and background needed to master many of the more complex but rewarding concepts contained within this versatile product. The book begins by introducing the basics of creating a project, manipulating data sources, choosing the right property values for each node, and navigating through different results windows. It then demonstrates various pre-processing tools required for building predictive models before beginning its treatment of the three main predictive modeling tools—Decision Tree, Neural Network, and Regression. These are addressed in considerable detail, with numerous examples o f practical business applications that are illustrated with tables, charts, displays, equations, and even manual calculations that let you see the essence of what Enterprise Miner is doing as it estimates or optimizes a given model. By the time you finish with this book, Enterprise Miner will no longer be a "black box": you will have an in-depth understanding of the product's inner workings. Throughout the book, the link between the output generated by Enterprise Miner and the statistical theory behind the business analysis for which Enterprise Miner is being used is explicitly shown. Even the SAS code generated by each node is examined line by line to show the correspondence between the theory and the results produced by Enterprise Miner. In many places, however, intuitive explanations are used to give you an appreciation of the way that various nodes such as Decision Tree, Neural Network, Regression, and Variable Selection operate and how different options such as Model Selection Criteria and Model Assessment are implemented. These explanations are intended not to replicate the exact steps that SAS uses internally to make these computations, but to give a good practical sense o f how these tools work in general. Overall, I believe this approach will help you use the tools in Enterprise Miner with greater comprehension and confidence. Several examples o f business questions drawn from the insurance and banking industries and based on simulated, but realistic, data are used to illustrate the Enterprise Miner tools. However, the procedures discussed are relevant for any industry. Tables and graphs from the output data sets created by various nodes are also included to give you an idea o f how to make custom tables from the results produced by Enterprise Miner. In the end you should have gained enough understanding o f Enterprise Miner to become comfortable and innovative in adapting the applications discussed here to solve your own business problems.

Chapter Details Chapter 1 discusses research strategy. I include general issues such as defining the target population, defining the target (or dependent) variable, collecting data, cleaning the data, and selecting an appropriate model.


Preface

Chapter 2 shows how to open Enterprise Miner, start a new project, and create data sources. It shows various components of the Enterprise Miner window and shows how to create a process flow diagram. In this chapter, I use example data sets to demonstrate in detail how to use the Input Data, StatExplore, MultiPIot, Impute, Data Partition, Filter, Variable Selection, Transform Variables, SAS Code, and Drop nodes. I also discuss the output and SAS code generated by some of these nodes. I manually compute certain statistics such as Cramer's V and compare the results with those produced by StatExplore. Chapter 3 covers the Variable Selection and Transform Variables nodes in detail. In using the Variable Selection node you have a choice of many options, depending on the type of target and the measurement scale of the inputs. To help make things clear, I illustrate each situation with a separate data set. Chapter 4 discusses decision trees and regression trees. First I present the general tree methodology, and, using a simple example, I manually work through the sequence of s t e p s — growing the tree, classifying the nodes, and pruning—whichis performed by the Decision Tree node. I then show how decision tree models are built for predicting response and risk by presenting two examples based on a hypothetical auto insurance company. The first model predicts the probability of response to a mail order campaign. The second model predicts risk as measured by claim frequency, and since claim frequency is measured as a continuous variable, the model built in this case is a regression tree. A detailed discussion of the SAS code generated by the Decision Tree node is included at the end of the chapter. Chapter 5 provides an introduction to neural networks. Here I try to demystify the neural networks methodology by giving an intuitive explanation using simple algebra. I show how to configure neural networks to be consistent with economic and statistical theory and how to interpret the results correctly. The neural network architecture—input layer, hidden layers, and output layer—is illustrated algebraicallywith numerical examples. Although the formulas presented here may look complex, they do not require a high-level knowledge of mathematics, and patience in working through them will be rewarded with a thorough understanding o f neural networks and their applications. In this chapter, the iterative processes of estimation of the model using the training data set as well as the selection of the optimal weights for the model using the validation data set are first discussed intuitively. Next, explicit numerical examples are given to clarify each step. As in Chapter 4, two models are developed using the hypothetical insurance data—aresponse model with a binary target, and a risk model with (in this case) an ordinal target representing accident frequency. I examine line by line the SAS code generated by the Neural Network node and show the correspondence between the theory and the results produced by Enterprise Miner. Chapter 6 demonstrates how to develop logistic regression models for targets with different measurement scales: binary, categorical with more than two categories, ordinal, and continuous (interval-scaled). Using an example data set with a binary target, I demonstrate various model selection criteria and model selection methods. I also present business applications from the banking industry involving two predictive models, one with a binary target and one with a continuous target. The model with a binary target predicts the probability of response to a mail campaign while the model with a continuous target predicts the increase in deposits that is due to an interest rate increase. This chapter also shows how to calculate the lift and capture rates of the models when the target is continuous.


Preface ix

In Chapter 7,1 compare the results of three modeling tools—DecisionTree, Neural Network, and Regression—that were presented in earlier chapters. For this purpose I develop two predictive models and then take turns applying the three modeling tools to each model. The first model has a binary target and predicts the probability of customer attrition for a fictitious bank. The second model has an ordinal target, which is a discrete version of a continuous variable, and predicts risk (as measured by loss frequency) for a fictitious auto insurance company. This chapter also provides a method of computing the lift and capture rates of these models using the expected value of the target variable. Chapter 8 shows how to calculate profitability for each of the ten deciles created when a data set of prospective customers is scored using the output of the modeling process. It then shows how to use these profitability estimates to address questions such as how to choose an optimum cut-off point for a mailing campaign. Here my objective is to introduce the notion of the marginal cost and marginal revenue associated with risk and response and to show how they can be used to make rational quantitative decisions in the marketing sphere.

How t o Use t h e Book •

To get the most out of this book, open Enterprise Miner and follow the sequence of tasks performed in each chapter, using either the data sets provided op this book's companion Web site or, even better, your own data sets. See support.sas.com/companionsites.

Work through the manual calculations as well as the mathematical derivations presented in the book to get an in-depth understanding of the logic behind different models.

To learn predictive modeling, read the general explanation and the theory, and then follow the steps given in the book to develop models using either the data sets provided on this book's companion Web site or your own data sets. Try variations of what is done in the book to strengthen your understanding of the topics covered.

I f you already know Enterprise Miner and want to get a good understanding of decision trees and neural networks, focus on the examples and detailed derivations given in Chapters 4, 5, and 6. These derivations are not as complex as they appear to be.

Prerequisites •

Elementary algebra and basic training (equivalent to one to two semesters of course work) in statistics covering inference, hypothesis testing, probability, and regression

Familiarity with measurement scales of variables—continuous, categorical,ordinal, etc.

Experience with Base SAS software and some understanding of simple SAS macros and macro variables


Acknowledgments I would like to thank a number o f people who helped me at various stages during the writing of this book. At the beginning of this project, I had many discussions about practical aspects of modeling techniques, especially neural networks and decision trees, with Ravindra Sarma, who was at that time a graduate student at MIT working on a neural networks project. In addition, he read the manuscript at several stages and contributed editorial assistance, as well as many in-depth questions and insightful comments that greatly improved the text. While I was working on SAS Enterprise Miner projects, I turned several times to SAS Technical Support. In addition to always answering my questions about how to run the software, they often provided informative explanations of the methodology behind it. I learned a great deal by sending questions to them and getting solid answers. I do not have a list of all the people I talked with in SAS Technical Support, so I would like to recognize them as a group. I would like to thank Randal Blank, a friend and a senior associate at KPMG LLP, for his constant encouragement and editorial help. Chris Monroe, a colleague from my days at A T & T , provided valuable editorial assistance on several sections. In addition, I would like to thank seven reviewers from SAS—Fang Chen, Brent Cohen,David Duling, Leonardo Auslender, Mary Grace Crissey, R. Wayne Thompson, and Rob Agnelli—and two outside reviewers, Goutam Chakraborty and Michael Szenberg. I could not have written this book without the encouragement and support of Julie Piatt and Patsy Poole at SAS Publishing. Julie was the supervisor of the project, and Patsy worked closely with me and was very helpful with organization and editing. Other members of the SAS Publishing team who made indispensable contributions were Caroline Brickley, who edited the book, Mary Beth Steinbach, who was the managing editor and reviewed the final version of the book, and Candy Farrell, the electronic production specialist whose wizardry was applied to the production of the book. I express my heartfelt thanks to all of them. I especially would like to thank my wife, Lokamatha Sarma, for putting up with me while I spent many weekends working on this book.


Research Strategy

1.1 1.2 1.3 1.4 1.5 1.6

Introduction Measurement Scales for Variables Defining the Target Sources of Modeling Data Pre-ProcessIng the Data Alternative Modeling Strategies

1 ..2 ....2 .......11 12 14

1.1 Introduction This chapter discusses the planning and organization of a predictive modeling project. Planning involves tasks such as these: •

defining and measuring the target variable in accordance with the business question

collecting the data

comparing the distributions of key variables between the modeling data set and the target population to verify that the sample adequately represents the target population

defining sampling weights if necessary

performing data-cleaning tasks that need to be done prior to launching SAS Enterprise Miner

Alternative strategies for developing predictive models using Enterprise Miner are discussed at the end of this chapter.


2

Predictive Modeling with SAS Enterprise Miner: Practical Solutions for Business Applications

1.2 Measurement Scales for Variables Because many of the steps above will involve a discussion of the data and types of variables in our data sets, it is important that I first define the measurement scales for variables that are used in this book. In general, I have tried to follow the definitions given by Alan Agresti': •

A categorical variable is one for which the measurement scale consists of a set o f categories.

Categorical variables for which levels (categories) do not have a natural ordering are called nominal.

Categorical variables that do have a natural ordering of their levels are called ordinal.

An interval variable is one that has numerical distances between any two levels of the scale.

According to the above definitions, the variables INCOME and AGE in Tables 1.1 to 1.5 and BAL_AFTER in Table 1.3 are interval-scaled variables. Because the variable RESP in Table 1.1 is categorical and has only two levels, it is called a binary variable. The variable LOSSFRQ in Table 1.2 is ordinal. (In Enterprise Miner you can change its measurement scale to interval, but I have left it as ordinal.) The variables PRIORPR and NXTPR in Table 1.5 are nominal. In some manuals and books, interval-scaled variables are called continuous. Continuous variables are usually treated as interval variables as a matter of practical convenience. Therefore I use the terms interval-scaled and continuous interchangeably. I also use the terms orderedpolychotomous variables and ordinal variables interchangeably. Similarly, I use the terms unordered polychotomous variables and nominal variables interchangeably.

1.3 Defining the Target The first step in any data mining project is to define and measure the target variable to be predicted by the model that emerges from your analysis of the data. This section presents examples of this step applied to five different business questions. The examples shown here are intended to highlight the fact that the tasks of defining and measuring the target variable can be nontrivial.

1.3.1 Predicting Response to Direct Mail In the example used in this book, a hypothetical auto insurance company wants to acquire customers through direct mail. The company wants to minimize mailing costs by targeting only the most responsive customers. Therefore, the company decides to use a response model. The target variable for this model will be RESP, and it is binary, taking the value of 1 for response and 0 for no response. Table 1.1 shows a simplified version of a data set used for modeling the binary target, response (RESP).

1

nd

Alan Agresti, Categorical Data Analysis, 2 ed. (New York, NY: John Wiley & Sons, 2002), 2.


Chapter 1: Research Strategy

3

Table 1.1 CUSTOMER

AGE

INCOME

1

23

2

43

3

34

4

32

S24.000

5

43

6

36

7

78

8

6

9

STATUS

PC

NC

RESP

S45.0Q0 S

1

1

0

S61.000

MC

1

2

:

MC

1

3

0

MNC

0

*

0

531,000 MC

0

5

0

323,436 MC

1

6

1

W

0

7

0

Si 00.25 6 D

1

1

0

26

S345.67S

MNC

1

: j

1

10

33

S100.2M

S

0

3

0

11

51

S21.312

MC

1

4

•

13

31

S83.456

0

5

!

13

23

524 J 34

MNC

1

1

0

14

47

543,566

MC

0

3

0

IS

77

512.002 MC

1

4

1

16

83

S32.454

W

1

5

:

17

23

561.345

S

0

6

0

18

32

S76.123

MC

1

7

0

19

52

S25.324

1

8

0

20

32

S31.3S6 MNC

0

1

0

21

23

S7S.345 S

1

8

0

22

80

S61.234

MNC

1

-

0

23

123

S76.S76

S

1

4

0

24

43

S24.002

3

5

0

In Table 1.1 the variables A G E , I N C O M E . STATUS, PC, and N C are input variables (or explanatory variables). A G E and I N C O M E are numeric and, although they could theoretically be considered continuous, it is simply more practical to treat them as interval variables. The variable S T A T U S is categorical and nominal-scaled. The categories o f this variable are S i f the customer is single and never married, M C i f married with children, M N C i f married without children, W i f widowed, and D i f divorced. The variable PC is numeric and binary. It indicates whether the customers own a personal computer or not, taking the value 1 i f they do and 0 i f not. The variable N C represents the number o f credit cards the customers own. You can decide whether this variable is ordinal or intervalscaled.


4 Predictive Modeling with SAS Enterprise Miner: Practical Solutions for Business Applications

The target variable is RESP and takes the value 1 if the customer responded, for example, to a mailing campaign, and 0 otherwise. A binary target can be either numeric or character; I could have recorded a response as Y instead of 1, and a non-response as N instead o f 0, with virtually no implications for the form of the final equation. Note that there are some extreme values in the table. For example, one customer's age is recorded as 6. This is obviously a recording error, and the age should be corrected to show the actual value, i f possible. INCOME has missing values that are shown as dots, while the nominal variable STATUS has missing values that are represented by blanks. The Impute node o f Enterprise Miner can be used to impute such missing values. See Chapters 2, 6, and 7 for details.

1.3.2

Predicting Risk in the Auto Insurance Industry

The auto insurance company wants to examine its customer data and classify its customers into different risk groups. The objective is to align the premiums it is charging with the risk rates of its customers. I f high-risk customers are charged low premiums, the loss ratios will be too high and the company will be driven out of business. I f low-risk customers are charged disproportionately high rates, then the company will lose customers to its competitors. By accurately assessing the risk profiles o f its customers, the company hopes to set customers' insurance premiums at an optimum level consistent with risk. A risk model is needed to assign a risk score to each existing customer. In a risk model, loss frequency can be used as the target variable. Loss frequency is calculated as the number o f losses due to accidents per car-year, where car-year is equal to the time since the auto insurance policy went into effect, expressed in years, multiplied by the number of cars covered by the policy. Loss frequency can be treated as either a continuous (interval-scaled) variable or a discrete (ordinal) variable that classifies each customer's losses into a limited number o f bins. (See Chapters 5 and 7 for details about bins.) For purposes o f illustration, I model loss frequency as a continuous variable in Chapter 4 and as a discrete ordinal variable in Chapters 5 and 7. The loss frequency considered here is the loss arising from an accident in which the customer was "at fault," so it could also be referred to as "at-fault accident frequency." I use loss frequency, claim frequency, and accident frequency interchangeably. Table l .2 shows what the modeling data set might look like for developing a model with loss frequency as an ordinal target.


Chapter 1: Research Strategy

Table 1.2 CUSTOMER 1

AGE

INCOME

N PR MO

25

545.000

0

D

45

561.000

1

:

^

0

S24.000

3

3

43

S31,000

4

3

6

56

523,456

0

0

7

73

1

8

6

S100.2S6

3

:

9

26

S345.67S

-

:

10

32

5100.211

5

3

11

51

S21.312

3

2

12

31

1

:

13

23

524.23-

0

3

14

47

543.566

1

0

15

77

S12.002

0

0

16

83

S32.454

0

2

17

25

561,343

1

:

576.123

1

0

3

54

4

32

5

IS 19

52

S25J24

3

:

20

32

S31.SS6

1

3

21

23

378,345

3

:

2?

80

S61.234

-

0

23

123

S76.S76

24

43

524.002

0 1

1


6 Predictive Modeling with SAS Enterprise Miner: Practical Solutions for Business Applications

The target variable is LOSSFRQ, which represents the accidents per car-year incurred by a customer over a period of time. This variable will be discussed in more detail in subsequent chapters in this book. For now it is sufficient to note that it is an ordinal variable that takes on values o f 0, 1,2, and 3. The input variables are AGE, INCOME, and NPRVIO. The variable NPRVIO represents the number of previous violations a customer had before he purchased the insurance policy.

1.3.3 Predicting Rate Sensitivity of Bank Deposit Products In order to assess customers' sensitivity to an increase in the interest rate on a savings account, a bank may conduct price tests. Suppose one such test involves offering a higher rate for a fixed period of time, called the promotion window. In order to assess customer sensitivity to a rate increase, it is possible to fit three types of models to the data generated by the experiment: •

a response model to predict the probability of response

•

a short-term demand model to predict the expected change in deposits during the promotion period

•

a long-term demand model to predict the increase in the level of deposits beyond the promotion period

The target variable for the response model is binary: response or no response. The target variable for the short-term demand model is the increase in savings deposits during the promotion period net o f any concomitant declines in other accounts. The target variable for the long-term demand model is the amount of the increase remaining in customers' bank accounts after the promotion period. In the case of this model, the promotion window for analysis has to be clearly defined, and only customer transactions that have occurred prior to the promotion window should be included as inputs in the modeling sample. Table l .3 shows what the data set looks like for modeling a continuous target.

2

If a customer increased savings deposits by $100 but decreased checking deposits by $20, then the net increase is $80. Here, net of means excluding.


Chapter h Research

Table 1.3 CUSTOMER

AGE

INCOME

B.JAN

B_FEB

B_MAR

B_APR

BAL_ AFTER

1

25

S45.000

54.000

S4.230

S4.400

54.900

55.900

2

45

S61.000

S5.G00

S4.000

S3.000

SO

52.000

3

54

51.200

SI.100

53.000

S100

S2O0

4

32

524.000

55.234

S345

35.678

S78

S87S

5

43

S31.000

S4.000

S4.230

54.400

S4.900

54.950

6

56

S23.456

52.000

S4.000

S3.000

SO

51,000

7

78

51.200

51.100

S3.CO0

5100

S1JO0

S

6

S100.256

S5.234

$345

55.678

S78

Sl.OSS

9

26

S345.67S

S3.435

S4.674

5678

SSO.00O

550.000

10

32

S100.211

S787

S4.230

S4.400

S4.900

55.900

ti

51

S21.312

SS.7S0

57.SOO

S3.456

SO

510.000

12

31

55.000

54.000

S3.000

so

S4,000

13

33

S24.234

54.000

54.230

S4.4Q0

S4.900

S5.900

14

47

S43.566

54,674

S678

SS00

S7,890

SS.S90

35

77

512,002

55,234

S345

S3.678

578

S1.07S

16

83

532.454

54,0X1

S4.230

S4.4O0

S4.900

S5.900

17

25

S61.343

S2.000

54.000

53.000

SO

SLOOO

18

33

S76.123

SI,200

SI. 100

53.000

Sioo

51.100

19

52

S25.324

S5.234

S345

55.678

S78

SLOTS

20

32

S31.SS6

33,435

34,674

S678

ss.ooo

S9.000

21

23

S7SJ45

$787

54.230

S4.4Q0

54.900

55.900

25

80

561.2 3-i

SS.7S0

57.SOO

53,456

SO

SIOO

23

123

576.876

S5.000

S4.000

S3.000

so

51,034

45

S24.0O2

S4.00Q

54.230

54.400

S4.900

S7,245

!

11

Strategy


8 Predictive Modeling with SAS Enterprise Miner: Practical Solutions for Business Applications

The data set shown in Table 1.3 represents an attempt by a hypothetical bank to induce its customers to increase their savings deposits by increasing the interest paid to them by a predetermined number o f basis points. This increased interest rate was offered (let us assume) in May 2006. Customer deposits were then recorded at the end of May 2006 and stored in the data set shown in Table 1.3 under the variable name BAL_AFTER. The bank would like to know what type of customer is likely to increase her savings balances the most in response to a future incentive of the same amount. The target variable for this is the dollar amount of change in balances from a point before the promotion period to a point after the promotion period. The target variable is continuous. The inputs, or explanatory variables, are AGE, INCOME, B_JAN, B_FEB, B MAR, and B_APR. The variables B_JAN, B_FEB, B_MAR, and B_APR refer to customers' balances in all their accounts at the end of January, February, March, and April of 2006, respectively.

1.3.4 Predicting Customer Attrition In banking, attrition may mean a customer closing a savings account, a checking account, or an investment account. In a model to predict attrition, the target variable can be either binary or continuous. For example, i f a bank wants to identify customers who are likely to terminate their accounts at any time within a pre-defined interval of time in the future, it is possible to model attrition as a binary target. However, if the bank is interested in predicting the specific time at which the customer is likely to "attrit," then it is better to model attrition as a continuous t a r g e t — time to attrition. In the example shown below, attrition is modeled as a binary target. When you model attrition using a binary target, you must define a performance window during which you observe the occurrence or non-occurrence of the event. I f a customer attrited during the performance window, the record will show 1 for the event and 0 otherwise. Any customer transactions (deposits, withdrawals, and transfers of funds) that are used as inputs for developing the model should take place during the period prior to the performance window. The inputs window during which the transactions are observed, the performance window during which the event is observed, and the operational lag, which is the time delay in acquiring the inputs, are discussed in detail in Chapter 7 where an attrition model is developed. Table 1.4 shows what the data set looks like for modeling customer attrition.


Chapter I ; Research Strategy

9

Table 1.4

| CUSTOMER

AGE

INCOME

B_JAN

B.FEB

B_MAR

B_.\PR

ATTR

1

25

545,000

34.000

S4.230

S4.400

S4.900

0

j

45

S61.000

35.000

S4.000

S3.000

SO

1

3

54

SI.200

S1.100

S3.000

S100

0

-

32

S24.000

SS.234

S345

S5.67S

S78

0

5

43

S31.000

S4.000

S4.230

S4.4O0

S4.900

0

6

56

S23.456

S2.000

S4.000

S3.000

SO

1

7

7S

SI.200

51,100

S3.000

S100

0

S

6

S10O.256

S5.234

S345

S5.67S

S78

0

9

26

S345.678

53.433

54,674

5673

SS0 000

1

10

32

S1002I1

S787

54,230

S4.40Q

54.900

0

11

5!

321.312

SS.7S0

S7.S00

S3.456

50

:

12

31

S5.000

S4.000

S3.000

SO

I

13

23

S24.234

S4.000

S4^30

54.400

54.900

0

M

47

S43.566

S4.674

S678

SSOO

S7.S90

0

15

77

SI 2.002

S5J34

S345

S5.67S

16

83

S32.454

S4.000

54,230

S4.400

S4.900

0

17

25

S61.345

S2.000

S4.000

S3.000

50

0

13

32

S76.123

SI. 200

SI.100

S3 000

5100

0

19

52

S25.324

S5.234

S345

S5.67S

S78

0

20

32

S31.SS6

S3.435

54.674

S678

SSO.000

0

21

23

S7S.345

S787

54.230

54.400

54.900

0

SO

S61.234

SS.7S0

S7.S00

S3.456

SO

0

23

123

S76.S76

S5.000

54.000

S 3.000

SO

0

24

45

S24.002

S4.000

S4.230

S4.4O0

S4.900

J

S78

1

In the data set shown in Table 1.4. the variable A T T R represents the customer attrition observed during the performance window, consisting o f the months o f June. July, and August o f 2006. The target variable takes the value o f ) i f a customer attrits during the performance window and 0 otherwise. Tablel.4 shows the input variables for the model. They are A G E . I N C O M E , B _ J A N . B_FEB, B M A R . and B_APR. The variables B J A N , B_FEB, B _ M A R , and B_APR refer to customers' balances for all o f their accounts at the end o f January, February. March, and A p r i l o f 2006, respectively.


Predictive Modeling with SAS Enterprise Miner: Practical Solutions for Business

Applications

1.3.5 Predicting a Nominal C a t e g o r i c a l (Unordered Polychotomous) Target Assume that a hypothetical bank wants to predict, based on the products a customer currently owns and other characteristics, which product the customer is likely to purchase next. For example, a customer may currently have a savings account and a checking account, and the bank would like to know if the customer is likely to open an investment account, or open an I R A , or take out a mortgage. The target variable for this situation is nominal. Models with nominal targets are also used by market researchers who need to understand consumer preferences for different products or brands. Chapter 6 shows some examples o f models with nominal targets. Table l .5 shows what a data set might look like for modeling a nominal categorical target. Table 1.5

CUSTOMER

AGE

INCOME

PRIORPR

NXTPR

1

25

S45.000

A

X

45

S61.000 B

z

3

54

c

V

4

32

S24.000

A

X

5

ÂŤ

S31.000 B

z

6

56

S23.456

C

z

7

78

C

z

S

6

S 100.256 A

X

9

26

$345,678

AB

X

10

32

Si 00,211 CD

z

11

51

13

31

13

23

14

S21.312

AC

Y

AB

X

524.234

CD

z

47

S43.566

D

z

15

77

S12.002 E

z

16

S3

532 ^-4 A

X

1?

25

S6I.345

B

X

IS

32

S76.123

A

z

19

52

S25J24

A

Y

20

32

S31,836

C

X

21

23

5-S.345

D

z

80

561.254

A

z

23

123

576.876

E

z

5*

45

S24.O02 D

X


Chapter 1: Research Strategy 11

In Table 1.5, the input data includes the variable PRIORPR which indicates the product or products owned by the customer of a hypothetical bank at the beginning of the performance window. The performance window, defined in the same way as in Section 1.3.4, is the time period during which a customer's purchases are observed. Given that a customer owned certain products at the beginning of the performance window, we observe the next product that the customer purchased during the performance window and indicate it by the variable NXTPR. For each customer, the value for the variable PRIORPR indicates the product that was owned by the customer at the beginning of the performance window. The letter A might stand for a savings account, B might stand for a certificate of deposit, etc. Similarly, the value for the variable NXTPR indicates the first product purchased by a customer during the performance window. For example, i f the customer owned product B at the beginning of the performance window and purchased products X and Z, in that order, during the performance window, then the variable NXTPR takes the value X. I f the customer purchased Z and X, in that order, the variable NXTPR takes the value Z, and the variable PRIORPR takes the value B on the customer's record.

1.4 Sources of Modeling Data It is important to distinguish between two different scenarios by which data becomes available for modeling. For example, consider a marketing campaign. In the first scenario, the data is based on an experiment carried out by conducting a marketing campaign on a well-designed sample o f customers drawn from the target population. In the second scenario, the data is a sample drawn from the results of a past marketing campaign and not from the target population. While the latter scenario is clearly less desirable, it is often necessary to make do with whatever data is available. In such cases, you can make some adjustments through observation weights to compensate for the lack of perfect compatibility between the modeling sample and the target population. In either case, for modeling purposes, the file with the marketing campaign results is appended to data on customer characteristics and customer transactions. Although transaction data is not always available, these tend to be key drivers for predicting the attrition event.

1.4.1 Comparability between the Sample and the Target Universe Before launching a modeling project it is necessary to verify that the sample is a good representation of the target universe. This can be done by comparing the distributions o f some key variables in the sample and the target universe. For example, i f the key characteristics are age and income, then you should compare the age and income distribution between the sample and the target universe.

1.4.2 Observation Weights I f the distributions o f key characteristics in the sample and the target population are different, sometimes observation weights are used to correct for any bias. In order to detect the difference between the target population and the sample, it is necessary to have some prior knowledge of the target population. Assuming that age and income are the key characteristics, you can derive the weights as follows: Divide income into, let's say, four groups and age into, say, three groups. Suppose that the target universe has N people in the /'* age group and / * income group, and 0

assume that the sample has n people in the same age-income group. In addition, suppose the i}

total number of people in the target population is TV, and the total number o f people in the sample


12 Predictive Modeling with SAS Enterprise Miner: Practical Solutions for Business Applications

is n. In this case, the appropriate observation weight is (Ny IN) /(n^ I n) for the individual in the /'* age group and / * income group in the sample. These observation weights should be constructed and included for each record in the modeling sample prior to launching Enterprise Miner, in effect creating an additional variable in your data set. In Enterprise Miner, you assign the role of Frequency to this variable in order for the modeling tools to consider these weights in estimating the models. The situation described here inevitably arises when you do not have a scientific sample drawn from the target population, which is very often the case. However, there is another source of bias that is often deliberately introduced. This bias is due to over-sampling of rare events. For example, in response modeling, i f the response rate is very low, it is necessary to include all the responders available and only a random fraction of nonresponders. The bias introduced by such over-sampling is corrected by adjusting the predicted probabilities with prior probabilities. This technique will be discussed in Section 4.7.2.

1.5 Pre-Processing the Data Pre-processing has several purposes: •

eliminate obviously irrelevant data elements, e.g., name, social security number, street address, etc., that clearly have no effect on the target variable

convert the data to an appropriate measurement scale, especially converting categorical (nominal-scaled) data to interval-scaled when appropriate

eliminate variables with highly skewed distributions

eliminate inputs which are really target variables disguised as inputs

impute missing values

Although many cleaning tasks can be done within Enterprise Miner, as the following examples show, there are some that should be done prior to launching Enterprise Miner.

1.5.1 Data Cleaning Before Launching SAS Enterprise Miner Data vendors sometimes treat interval-scaled variables, such as birth date or income, as character variables. I f a variable such as birth date is entered as a character variable, it would be treated by Enterprise Miner as a categorical variable with many categories. To avoid such a situation, it would be better to derive a numeric variable from the character variable and then drop the original character variable from your data set. For example, in one modeling project, I first converted the birth date to SAS format, derived age from the birth date, and then used age as an input variable. Similarly, income is sometimes represented as a character variable. The character A may stand for $20K ($20,000), B for $30K, etc. To convert the income variable to an ordinal or interval scale, it is best to create a new version of the income variable in which all the values are numeric, and then eliminate the character version of income. Another situation which requires data cleaning that cannot be done within Enterprise Miner arises when the target variable is disguised as an input variable. For example, a financial institution would like to model customer attrition in its brokerage accounts. A model is to be developed to predict the probability o f attrition during a time interval of three months in the future. The institution decides to develop the model based on actual attrition during a performance window of


Chapter 1: Research Strategy

13

three months. The objective is to predict attritions based on customers" demographic and income profiles and balance activity in their brokerage accounts prior to the window. The binary target variable takes the value o f 1 i f the customer attrits and 0 otherwise. I f a customer's balance in his brokerage account is 0 for two consecutive months, then he is considered an atlritor. and the target value is set to I . I f the data set includes both the target variable (attrition/no attrition) and the balances during the performance window, then the account balances may be inadvertently treated as input variables. To prevent this, inputs which are really target variables disguised as input variables should be removed before launching Enterprise Miner. These examples demonstrate that while there is no uniformly applicable solution to data cleaning, every effort should be made to ensure that instances such as those illustrated above are resolved before launching Enterprise Miner.

1.5.2

Data C l e a n i n g After L a u n c h i n g S A S E n t e r p r i s e Miner

Display 1.1 shows an example o f a variable that is highly skewed. The variable is M S , which indicates the marital status o f a customer. The variable RESP represents customer response to mail. It takes the value o f 1 i f a customer responds, and 0 otherwise. In this hypothetical sample, there are only 100 customers with marital status M (married), and 2900 with S (single). None o f the married customers are responders. An unusual situation such as this may cause the marital status variable to play a much more significant role in the predictive model than is really warranted, because the model tends to infer that all the married customers were non-responders because they were married. The real reason there were no responders among them is simply that there were so few married customers in the sample. Display 1.1

The SAS System The FREQ Frequency Percent Row Pet Col Pet

Procedure

Table of MS by R E S P RESP(RESPONSE) MS(MARITAL STATUS)

0

1

Total

M

100 3.33 100.00 3.42

0 0.00 0.00 0.00

100 333

2B26 94.20 97.45 96.58

74 2.47 2.55 100.00

2900 96.67

2926 97.53

74 2.47

3000 100.00

s

Total

Variables such as the one shown in Display I . I can produce spurious results i f used in the model. You can identify these variables using the StatExplore node, set their roles to Rejected in the Input Data node, and drop them from the table using the Drop node.


14 Predictive Modeling with SAS Enterprise Miner: Practical Solutions for Business Applications

The Filter node can be used for eliminating observations with extreme values, although I do not recommend elimination of observations. Correcting them or capping them instead may be a better alternative, in order to avoid introducing any bias into the model parameters. The Impute node offers a variety of methods for imputing missing values. These nodes will be discussed in detail in the next chapter. Imputing missing values is necessary when you use the Regression or Neural Network nodes.

1.6 Alternative Modeling Strategies The choice of modeling strategy depends on the modeling tool and the number o f inputs under consideration for modeling. Here are examples of two possible strategies to consider in using the Regression node.

1.6.1 Regression with a Moderate Number of Input Variables Pre-process the data: •

Eliminate obviously irrelevant variables.

Convert nominal-scaled inputs with too many levels to numeric interval-scaled inputs, i f appropriate.

Create composite variables (such as average balance in a savings account during the six months prior to a promotion campaign) from the original variables i f necessary. This can also be done with Enterprise Miner using the SAS Code node.

Next, use Enterprise Miner to perform these tasks: •

Impute missing values.

Transform the input variables.

Partition the modeling data set into train, validate, and test (when the available data is large enough) samples. Partitioning can be done prior to imputation and transformation, because Enterprise Miner automatically applies these to all parts of the data.

Run the Regression node with the Stepwise option.

1.6.2 Regression with a Large Number of Input Variables Pre-process the data: •

Eliminate obviously irrelevant variables.

Convert nominal-scaled inputs with too many levels to numeric interval-scaled inputs, i f appropriate.

Combine variables i f necessary.

Next, use Enterprise Miner to perform these tasks: •

Impute missing values.

Make a preliminary variable selection. (Note: This step is not included in Section l .6.1.)

Group categorical variables (collapse levels).


Chapter 1: Research Strategy 15

Transform interval-scaled inputs.

Partition the data set into train, validate, and test samples.

Run the Regression node with the Stepwise option.

The main difference between the steps outlined in Sections 1.6.1 and 1.6.2 is that Enterprise Miner is used to make a preliminary variable selection in Section 1.6.2 (because the number o f inputs is large) and not in Section 1.6.1. The steps given in Sections 1.6.1 and 1.6.2 are only two of many possibilities. For example, one can use the Decision Tree node to make a variable selection and create dummy variables to then use in the Regression node.


Getting Started w i t h Predictive Modeling

2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8

Introduction Opening SAS Enterprise Miner 5.2 Creating a New Project in SAS Enterprise Miner 5.2 The SAS Enterprise Miner Window Creating a SAS Data Source Creating a Process Flow Diagram Summary... Appendix to Chapter 2

17 18 20 ...21 22 35 ....72 72

2.1 Introduction This chapter introduces you to S A S Enterprise Miner 5.2 and some of the preprocessing and data cleaning tools (nodes) needed for data mining and predictive modeling projects. The following topics are covered: •

opening Enterprise Miner 5.2

starting a new project:

o

naming the project

o

providing a directory where Enterprise Miner stores the data sets and programs that it generates for the project

o

specifying the directory where the user's data sets are located.

creating a data source for each data set used in the project: o

providing the name and location of the data set

o

setting the measurement scales and roles of the input variables

o

creating a profit matrix for decision-making.


18

Predictive Modeling with SAS Enterprise Miner: Practical Solutions for Business Applications

•

using some of the nodes for data cleaning, data exploration, variable selection, and variable transformation: o

Input Data node (for specifying the data set for the project)

o

StatExplore node (for univariate analysis of the inputs and exploring the relationship of inputs to the target)

o

MultiPlot node (also for univariate analysis of the inputs and exploring the relationship of inputs to the target)

o

Impute node (for imputing missing values)

o

Data Partition node (for partitioning the data set into subsets for use in training, validating, and testing models)

o

Filter node (for examining and filtering outliers)

o

Variable Selection node (for selecting important variables for modeling)

o

Transform Variables node (for performing various transformations to improve model accuracy)

o

S A S Code node (for writing custom code)

o

Drop node (for dropping variables).

Note: The order in which the nodes are presented is not necessarily the order in which they should generally be used. Enterprise Miner's modeling tools—Decision T r e node, e Neural Network node, and Regression node—are not included in this chapter as they are covered extensively in Chapters 4, 5, and 6.

2.2 Opening SAS Enterprise Miner 5.2 1

To start Enterprise Miner 5.2, click on the Enterprise Miner icon on your desktop . The logon screen will appear as shown in Display 2.1.

1

All the illustrations presented here are from sessions in which Enterprise Miner 5.2 is installed on a personal computer.


Chapter 2: Getting Started with Predictive Modeling

19

Display 2.1

1

SAS Analytics Platform

§sa& SAS' Enterprise Miner 5.2 The

ftmtr

to Know

Copyright @ 2 0 0 2 - 2 0 0 5 by SAS Institute I n c . C»ry,MC. USA. A l l Rights Reserved,

User name:

^asadm

Password

0

Remember my password

0

Personal Workstation

Log On

Cancel

Enter your user name and password in the logon screen. Check the Personal Workstation box i f you are using Enterprise Miner 5.2 on a personal computer. When you click the Log O n button, the Enterprise Miner w i n d o w opens, as shown in Display 2.2. Display 2.2 Si Enterprise Miner fit

Edt

v*w

Actions Options Window Help W e l c o m e to Enterprise Miner

Enterprise Miner

1

;*S

V'

Help Topics

'ip

New Project...

—' -

• B P

Open Project.. Recent Projects


20

Predictive Modeling with SAS Enterprise Miner: Practical Solutions for Business Applications

2.3 C r e a t i n g a N e w P r o j e c t in SAS E n t e r p r i s e M i n e r 5 . 2 Creating a new project involves naming the project, as well as indicating the directory where you want the project to be saved and the directory where the data for the project exists. These steps are shown below. When you select New Project from the Enterprise Miner window shown in Display 2.2, the Create New Project window opens, as shown in Display 2.3. Display 2.3 Create Mew Project General | Start-Up Code j Exit Code Name: chapter2

Host:

SASMain - Logical Workspace Server

Path;

C:\TheBook'iEM 5.2Wov2006

3

OK

Cancel

|

Help

In this window enter the Name o f the project and the Path where you want to save the project. I typed chapter2 in the Name box, and typed C: \TheBook\EM_5 . 2\Nov2 006 in the Path box. This is the directory where the project w i l l be stored. Next, I selected the Start-Up Code tab. On the Start-Up Code tab, I typed the L I B N A M E statement as shown in Display 2.4. Display 2.4 Create New Project: General Slart-Up Code | Exit Code | 1 2 3

o p t i o n s nofrnterr ; lib-name TlieBook. "C: \Tlie3ook\8ata"

;

The data for this project is located in the folder c: \TheBook\Data. This w i l l be indicated by the libref TheBook. When you select O K , the Enterprise Miner 5.2 interface window opens, showing the new project. This window is shown in Display 2.5.


t 'hapter 2: Gelling Storied with Predictive Modeling

21

2.4 The SAS E n t e r p r i s e M i n e r W i n d o w This is the window where you create the process flow diagram for your data mining project. In Display 2.5,1 have numbered the components o f the Enterprise Miner window for easy reference. Display 2.5 Enterprise M i r e r - C h a p t e r 2 File

O

Edit

View

Actions

Options

Window

Help

QIJBIH fc|

© ©

0

Sample] Explore j Motify | Mqd"ejJ Assess | Ulilty f

5fe a • 13

• Stiff

13 Chapter2 * Q Data Sources - J^J Diagrams

o

Response 9i Model Packages

it

Property

©

Value EMVys Response Available

Name Status

ID Diagram Identifier. This identifier

0

corresponds to the SAS libref used to identify the physical location of the contents of this diagram on the server

' * Connected to SASMain - Logical Workspace Serve'

Diagram Response opened

A brief description o f these components follows: O Menu Bar: The menu bar has seven items. These are File, Edit. View. Actions. Options. Window, and Help. By clicking on any one o f these buttons, you w i l l see a drop-dow n menu. For example, by clicking on View-MVopcrty S h e e t s Advanced, you can reset the properties list from Basic to Advanced. This is illustrated in Display 2.6. Display 2.6 Enterprise M i n e r - C h a p t e r 2 File •

Ed.t

View

Actions

Options

j Window

Property Sheet

Dii

Sample j J • * QDat * _4]Dia. •

4JM

-

BJUH

W

1

[*] Program Editor

Ctrl+Alt+P

Q

Log

Ctrl+Alt+L

O Output

Ctrl+Alt+O

Graphs

Ctrl+fllc+G

Table...

Help

Ctrl+Alt+T

Refresh Project Tree Ctrl+Alt+R

Basic • Advanced Hide


22

Predictive Modeling with SAS Enterprise Miner: Practical Solutions for Business Applications

©

Toolbar: The toolbar contains the Enterprise Miner node (tool) icons. The icons displayed on the toolbar change according the Node (Tool) G r o u p tab you select in the next line indicated by © in Display 2.5.

© Node (Tool) Groups: These tabs are for selecting different groups of nodes. The node groups are Sample, Explore, Modify, Model, Assess, and Utility. The toolbar © changes according to the node group selected. If you select the Sample tab on this line, you will see the icons for Input Data, Sample, Data Partition, and T i m e Series in ©. I f you select the Explore tab, you will see the icons for Association, Cluster, Multiplot, Path Analysis, S O M / K o h o n e n , StatExplore, Text Miner, and Variable Selection in ©. Similarly, other tabs can be used according to the tasks you are performing in Enterprise Miner. ©

ProjectPanel: The project panel is for viewing, creating, deleting, and modifying the Data Sources, Diagrams, Model Packages, and Users. For example, if you want to create a data source (that is, tell Enterprise Miner where your data is and give information about the variables, etc.), you click on Data Sources and proceed. For creating a new diagram, you right-click on Diagrams and proceed. To open an existing diagram, double-click on the diagram you want.

©

Properties Panel: The properties ofProject, Data Sources, Diagrams, Nodes, Model Packages, and Users are shown in this panel. Note that, in our example, the nodes are not yet created; hence, you do not see them in Display 2.5. Y o u can view and edit the properties of any object selected. I f you want to specify or change any options in a node such as Decision T r e e or Neural Network, you must use the Properties Panel.

©

Help Panel: This displays a description of the property that you select in the Properties Panel.

©

Status B a r : This indicates the execution status of the Enterprise Miner task.

©

Toolbar Shortcut Buttons: These are shortcut buttons for Create Data Source, Create Diagram, R u n , etc. To display the text name of these buttons, position the mouse pointer over the button.

©

Diagram Workspace:This is used for building and running the process flow diagram for the project with various nodes (tools) of Enterprise Miner.

2.5 Creating a SAS Data Source Y o u need to create a data source for each data set you use in the project. Creating a data source means providing Enterprise Miner with the relevant information about the data set being imported—the dataset's name, location, library path, the role assignments and measurement levels of its variables, and the decision variables attributed to the target such as the profit matrix, etc. The profit matrix, also referred to as Decision Weights in Enterprise Miner 5.2, is used in decisions such as assigning a target class to an observation and assessing the models. (The use of profit matrices is discussed in detail in Chapter 4.) Enterprise Miner saves all of this information, or metadata, as different data sets in a folder called DataSources in the Project Directory. To create a data source, click on the toolbar shortcut button or right-click on Data Sources in the Project Panel, as shown in Display 2.7.


Chapter 2: Getting Started with Predictive Modeling

23

Display 2.7 Enterprise Miner - Chapter2

g j Mj

n

BQ

®m

•B

ill

£33

Sample ! E>J)!ore ; Modify : Model i Assess Utility 13 :iiapter2

Diagrams f £ ] Create Data Source • • i»J Mgdd Packages =y Property

Name Project Name

I * , Connected to SASMah - Logical VNtorkspace Server

When you click on Create Data Source, the Data Source Wizard window opens, and Enterprise Miner prompts you to enter the data source. This window is shown in Display 2.8. Display 2.8

•HE)

$ Data Source Wizard - Step 1 of B Metadata Source

Select a metadata source

Source: S A S Table

Next>

Cancel

H e I i >

. I

Is the source a S A S Table or is it a Metadata Repository? If, as in my example, you are using a SAS data set in your project, enter S A S Table in the Source box and click on Next>. S A S Table is, by default, already selected in the Source box. So, you can simply click on Next>. Then another window opens, prompting you to give the location o f the SAS data set. This window is shown in Display 2.9.


24

Predictive Modeling with SAS Enterprise Miner: Practical Sol it! ions for Business Applications

Display 2.9

•IS

Data Source Wizard - Step 2 of 6 Select a SAS Table

J

.....

Select a SAS table

Browse

Table:

4

< Back

)(

Next >

|

|

Cancel

j

[

Help

When you click Browse.., a window opens that shows the list o f library references. This window is shown in Display 2.10. Display 2.10 mm

Select a SAS Table .i) 9 J ||

MAPS SAMPSIO SASHELP THEBOOK

Cancel

Since the data for my example project is in the library TheBook, when I double-click on T H E B O O K , the window opens with a list o f all the data sets in that library. This window is shown in Display 2.11.


Chapter 2: Getting Started with Predictive Modeling

25

Display 2.11 Select a SAS Table NEURAL NEURALDATA MEWTAL NUMRIJ3TARG NUMRI_NTARG PRDSALE PR ICE TEST PRICETEST_SCORE RESP

RESPB RESPB1 RESPB2 RESPB3 RESP_MOD RESP.SMPL. RESP_SMPL._CL.EAN RESP_SMPL_CLEAN2 RESP_SMPLâ&#x20AC;&#x17E;CLEAN3 RISK Properties.

OK

Cancel

I select the data set named N N _ R E S P _ D A T A . When 1 click O K , the Data Source Wizard window opens as shown in Display 2.12. Display 2.12 Data Source Wizard

- Step ? of 6 Select a SAS Table

Select a SAS table

Table:

Browse..

THEBOOK NN RESP DATA

= Back

Next>

Cancel

He*

This display shows the libref and the data set name, which I entered. When I press Next>. another window opens displaying the Table Properties. This is shown in Display 2.13.


Predictive Modeling with SAS Enterprise Miner: Practical Solutions for Business Applications

Display 2.13 Y D a l a S o u r c c Wizard ~ S t e p 3 o f 6 Table Information

Table Properties Value THE BOOK NN RESP DATA

Property Table Name Description Member Type Data Set Type Engine Number ol Variables Number ol Observations Created Date wod it ted Dote

DATA DATA V9 18 29904 2005-03-11 0B:S052.015 2005-03-11 08:50:52.015

Help

Clicking on Next> opens another window that shows Metadata Advisor Options. This is shown in Display 2.14A. Display 2.14A .,Y Data Source Wizard -- Step 4 of 6 Metadata Advisor Options

Metadata Advisor Options Use trie basic setting to set (he initial measurement levels and roles based on the variable attributes. Use the advanced setting to set the Initial measurement levels and roles based on both the variable attributes and distributions.

O Basic

[

@ Advanced

-Back

|| Next*

Customize

|

( Cancel

| | Help j

Display 2.I4A provides a snapshot o f the Metadata Advisor Options window, which is used for defining the metadata. Metadata is data about data sets. It specifies how each variable is used in the modeling process. The metadata contains information about the role o f each variable, its measurement scale, etc. I f you select the Basic option, the initial measurement levels and roles are based on the variable attributes. This means that i f a variable is numeric, its measurement scale is designated to be interval, irrespective o f how many distinct levels the variable may have. For example, a numeric binary variable w i l l initially be given the interval scale. I f your target variable is binary in numeric form, it w i l l be treated as an interval-scaled variable, and it w i l l be treated as such in the subsequent nodes. I f the subsequent node is a Regression node. Enterprise Miner automatically


Chapter 2: Getting Started with Predictive Modeling

27

uses ordinary least squares regression, instead o f the logistic regression, which is usually appropriate with a binary target variable. By selecting the Basic option, all character variables w i l l be considered nominal. You may reset the measurement scales and roles as appropriate. By selecting the Advanced option. Enterprise Miner applies a bit more logic as it automatically sets the variable roles and measurement scales. I f a variable is numeric and has more than 20 distinct values, Enterprise Miner sets its measurement scale to interval. In addition, i f you select the Advanced option, you can customize the measurement scales. For example, by default the Advanced option sets the measurement scale o f any numeric variable to nominal i f it takes 20 or fewer unique values, but you can change this number by clicking the Customize button. Display 2 . I 4 B shows the default settings for the Advanced Advisor Options. Display 2.14B t-, A d v a n c e d Advisor O p t i o n s

[x]|

:

I

Prooertv Value Missing Percentage Threshold 50 RejectVars with Excessive Missing Value Yes Class Levels Count Threshold 20 DelectClass Levels Yes Relect Levels Count Threshold fob ReiectVars with Excessive Class Values Yes Class L e v e l s C o u n t T h r e s h o l d I f "Detect class levels''~Yes, interval variables w i t h less than the number specified for this property w i l l be marked as N O M I N A L . The default value is 20

]

OK

j |

Cancel

j [

Help

|

One advantage o f selecting the Advanced option is that Enterprise Miner w i l l automatically set the role i f each unary variable to Rejected. I f any o f the settings are not appropriate, you can change them later in the window shown in Display 2.15. In this example, l changed the Class Levels Count Threshold property to 5 and closed the Advanced A d v i s o r Options window by clicking on O K and then clicking on Next>. This brings up the window shown in Display 2.15.


Predictive Modeling with SAS Enterprise Miner: Practical Solutions for Business Applications

Display 2.15 : Data Source Wizard

-Step 5 of 6 Column Metadata Show code Explore Name APE CRED DEUNO DEPC EMP STA GENDER HEQ INCOME M.FDU MILEAGE MOB MRTGI MS NUMTR RESTYPE RES_STA cuscode resp

j

Role

Input Input Input Input Input Input Input Input Input Input Input Input Input Input Input Input Rejected Input

Report

Level nterval nterval Nominal binary Nominal Binary Nominal Nominal Binary nterval Binary Nominal Nominal Nominal Nominal Elinary Nominal v Binary

Order

No NO No No No No No No No No No No No No No No No No

h.

Drop NO No No Ho Mo No No No Ho Ho NO No No No No No No NO

Lower

I

Help

In this window, a tabic is shown with the variable names, model roles, and measurement levels o f the variables in the data set. For the purpose o f this illustration, 1 specify the model role o f the variable RESP as the target, and click Next>. Enterprise Miner recognizes that the target is binary. Hence, it opens another window with the question Do you want to build models based on the values of the decisions? This window is shown in Display 2.16. Display 2.16 •jData Source Wizard ~ Step 6 of 0 Decision Crmllijut alton

Decision Processing Do von want to build models based on the values ol the decisions ? II you answer yes. you may enter Information on the cost or protS of each possible decision, prior probability and cost lunction. The data will be scanned tor the distributions ol the target variables.

r

NO

< Back

t?

Next >

Yes

Cancel

Help

I f you are using a profit matrix (decision weights), cost variables, and posterior probabilities, you should select Yes, and click Next > to enter these values. (These matrices can also be entered and modified at later stages.)


Chapter 2: Getting Started with Predictive Modeling

29

A window opens with tabs for Targets. Prior Probabilities. Decisions, and Decision Weights. The window for the Targets tab is shown in Display 2.17. Display 2.17 : Data Saurte Wizard â&#x20AC;&#x201D;Step 7 of 6 Decision Cniifiauraiioo Targets | prior PiobabWies [ Decisions | Decision Weights | Name :

resp

Measurement Level

Binary

Target level order:

Descending

Event level:

I

Format:

Help

The Targets tab displays the name o f the target variable and its measurement level. It also gives the target levels o f interest. In this example, the variable RESP is the target and is binary, which means that it has two levels, namely response, indicated by I , and non-response, indicated by 0. The event o f interest is response. That is, the model is set up to estimate the probability o f response. I f the target has more than two levels, this window w i l l show all o f its levels. In later chapters, I w i l l model an ordinal target that has more than two levels, each level indicating frequency o f losses or accidents, with 0 indicating no accidents, I indicating one accident, and 2 indicating two accidents, and so on. Display 2.18 shows the Prior Probabilities tab. Display 2.18 : Data Source Wizard â&#x20AC;&#x201D; S t e p 7 o F G

H

Targets Prior ProbaUiWies j Decisions | Decision Weights | Do you wont to errter new prior probabilities? F Ye* T No Level I 0

Court ... 10525

Pr-i 0.3i36 0.6854

Aci.urtfJFTi.jr 0.03 0 97

Help


30

Predictive Modeling with SAS Enterprise Miner: Practical Solutions for Business Applications

This tab shows (in the column labeled Prior) the probabilities o f response and non-response calculated by Enterprise Miner for the sample used for model development. In the modeling sample I used in this example, the responders are over-represented. In the sample there are 31.36% responders and 68.64% non-responders. These are called prior probabilities. So the models developed from the modeling sample at hand w i l l be biased unless a correction is made for the bias caused by over-representation o f the responders. I f you enter these prior probabilities in the Adjusted Priors column as I have done above, Enterprise Miner w i l l correct the models for the bias and produce unbiased predictions. To enter these adjusted prior probabilities, select Yes in response to the question Do you want to enter new prior probabilities? Then enter the probabilities that you calculated separately for the entire population (.03 and .97 in my example). In order to enter a profit matrix, click on the Decision Weights tab. Display 2.19 shows the Decision Weights (profit matrix) window. Display 2.19 -'.[Data Source Wizard -- Step 7 of 8 Decision Configuration Targets | Prior Probabilities | Decisions Decision Weights Select a decision function: C Minimize

C MaxMxe

Eriter weiglit volues for the decisions DECISION2 | DECISION! Level jog ...|12.0 1 J 00 _j_1i,g"_

< Back

Next*

Cancel

Help

The columns o f this matrix refer to the different decisions that need to be made based on the model's predictions. In this example, Dccisionl means classifying or labeling a customer as a responder, while Decision2 means classifying a customer as a non-responder. The entries in the matrix indicate the profit or loss associated with a correct or incorrect assignment, so the matrix in this example implies that i f a customer is classified as a responder, and he is in fact a responder. then the profit is $12. I f a customer is classified as a responder, but she is in fact a non-responder, then there w i l l be a loss o f $1. Other cells o f the matrix can be interpreted similarly. In developing predictive models, Enterprise Miner assigns target levels to the records in a data set. In the case o f a response model, assigning target levels to the records means classifying each customer as a responder or non-responder. In a later step o f any modeling project, Enterprise Miner also compares different models, on the basis o f a user-supplied criterion, to select the best model. In order to have Enterprise Miner use the criterion o f profit maximization when assigning target levels to the records in a data set and when choosing among competing models, select Maximize for the option Select a decision function. The values in the matrix shown here are arbitrary, and given only for illustration. Display 2.20 shows the window for cost variables.


Chapter 2; Getting Started with Predictive Modeling 31

Display 2.20 Data Source Wizard —Step 7 of a Decision Co nfigurat ion Targets | Prior Prctoabilities Decisions | Decision Weights | Do you want to usa the decisions? C No

P Yes Decision Nomoj DECISION1 Jl DECISION2 ]6

Label

| Cost Variable Constant i«None » 00 < None > 00

j

Help

I f you want to maximize profit instead o f minimizing cost, then there is no need to enter cost variables. Costs are already taken into account in profits. Therefore, in this example, cost variables are not entered. When you click on Next>, another window opens showing the data set and its Role. Display 2.21 shows this window. In the display, the data set N N _ R E S P _ D A T A is being assigned a role o f Raw. Display 2.21 Data Source. Wizard --Step If of 8 Data Source Attribute*

You may change the name and the tole, and can specify a population segment identifier for the data source to be created.

Name:

|NN_RESP_DATA

Role: ^

[ROVV

" 5

Segment: (~

Noles:

Fnsh

Cancel

Help


Predictive Modeling with SAS Enterprise Miner: Practical Solutions for Business Applications

Other options for Role are T r a i n , Validate, Test, Score, Document, and Transaction. Since I plan to create the T r a i n , Validate, and Test data sets from the sample data set, 1 leave its role as Raw. When 1 click on Finish, the Data Source Wizard closes, and the project window opens as shown in Display 2.22. Display 2.22

I have upgraded the Property Sheet from Basic to Advanced by clicking on View-> Property Shcet ^Advanced, as shown in Display 2.23. -

Display 2.23 _ ! • ; x; Fie Edrt View Actions Options V. Enterprise Miner - chapter? _

Window Help Baste

[*] Program Ector Qrl+Alt+P

S*3>!eJ 23 chapter

HLOQ Q Output

Orr+Afc+t Ctrl+AH+O

ffi Q D a t Graph, • J^jDiai— • yjMor Table...

Ctrl+AH+G

•*: jlJUs<

Ctrl+Afc+R

E

Refresh Project

Properly

1

if wuvair.eu

H i m ® D\D\m ,\

9 Isia

m &|

Ctrl+Aft+T

Value

f i * . Connected to SASMain - Logical Workspace Server

Display 2.24 shows Advanced properties o f the project named Chapter2.


Chapter 2: Getting Started with Predictive Modeling

33

Display 2.24 • Enterprise Miner-Chaplcr2 1 Fie

Edi

Swrpio *

View

Actions

E'f^e

Optmt

Window

neb

. Model - * w e « U « y

-' 1

SiH^

Q D s I a Sources • J (Wdel BaOirioe:

Vl*» Name

sun-up cod* Ect code

Server Grid Available

S A S N a i n - LooicalYVorVspacB Server No C l T h e B o o i d E M _ 5 TlNcrviQOBlChapWii

Max C o n c u r r a n l T a s k s

Default

HUM f l * , Connected io SASMan • L f f j c M Viat Jooce Serve.

In order to see the properties o f the data set, expand Data Sources in the project panel by clicking on the plus sign located to the left o f Data Sources. Then when you select the data set N N _ R E S P _ D A T A . the properties panel shows the properties o f the data source as shown in Display 2.25. You can view and edit the properties. You can also see a list o f variables by clicking on located on the right o f the Variables property in the Value column. The arrow in Display 2.25 shows where to click the mouse pointer, and Display 2.26 shows the window that opens as a result.


Predictive Modeling with SAS Enterprise Miner: Practical Solutions for Business Applications

Display 2.25 ^Enterprise Miner - chapter 2 File Edit View Actions Options Window Help ° E L*3 0

D

Sample 1 Explore j Modify j Model j Assess | Utility | 3 cnapter2 - n ] Data Sources NN RESP DATA

• H Diagrams Model Packages JTJ Users Property ID Name

NNRESPDATA NN RESP DATA

Variables

Decisions

pole

Raw

JSegme.nt .External Data Sampling Rate Run Code Notes

Yes

THEBOOK NN_RESP_DATA DATA 29904.0 Tie kssarmatsasadm 10/3/05 12:46 PM sasadm 10/4/05 6:17 AM Local

Library Table Tffit No.Obs.

No. Cols.

Created By Create Date Modified By Modify Date Scope

ID Data Source identifier The metadata tables are stored in the SAS library, and use this identifier as its LIBREF. Connected to SASMain - Logical Workspace Server

Display 2.26

rttn*

Rote input

CRED DELINO

Input

DEPC

Input Input

UFPlj

rnpul

EMP__STA Input LENDER Input hiptit HEG INCOME Input

Input input 'injujl MS Input rJUMTR Input RESTYFE Input R E S . S T A Input custrjde Refected rasp Taryel MrlEAOE MOB MRTC-I

Outer

Leva I':1s IV. 11 Interval Nominal Binary Norniial Binary nominal Nominal Binary Interval Brnary

jiorntail

Nominal Nominal Nominal Binary Nominal Binary

[_ r>oc

Ho

lower U r *

1

UapaLmt

j

type

-L

No

IS m •i EdtlBma SAS Coo*

This window shows the name, role, measurement scale, etc., for each variable in the data set. Y o u can change the role o f any variable, change its measurement scale, or drop a variable from the data set. I f a variable is dropped from the data set, it w i l l not be available in any o f the subsequent nodes.


Chapter 2: Getting Started with Predictive Modeling

35

2.6 C r e a t i n g a P r o c e s s Flow Diagram To create a process flow diagram, right-click on Diagrams in the project panel (shown in Display 2.25) and click on Create Diagram. Y o u w i l l be prompted to enter the name o f the diagram in a text box labeled Diagram Name. After entering a name for your diagram, click O K . Then a blank workspace opens, as shown in Display 2.27A, where you create your process flow diagram. Display 2.27A a

m

ffi b

' SsmrJs • E>r*ie Mottly Uo*l Amis Lilly 3 Chr*(«2 RESPONSE Property ID

Hsme

2.6.1

vnta EMWS RESPONSE Op»n

Input Data Node

This is the first node in any diagram (unless you start with the SAS Code node). In this node, you specify the data set that y o u want to use in the diagram. There may be several data sources that you may have already created for this project, as discussed in Section 2.5. From these data sources, you need to select one for this diagram. A data set can be assigned to an input in one o f two ways: •

When you expand Data Sources in the project panel by clicking on the [+] on the left o f Data Sources, all the data sources w i l l appear. Then click on the icon to the left o f the data set you want, and drag it to the Diagram Workspace. This w i l l create the Input Data node with the desired data set assigned to it.

Alternatively, first drag the Input Data node from the toolbar into the Diagram Workspace. Then set the Data Source property o f the Input Data node to the name o f the data set. To do this, select the Input Data node, then click on Q

located to the right o f

the Data Source property as shown in Display 2.27B. The Select Data Source window opens, as shown in Display 2.28. Then click on the data set you want to use in the diagram. Then click O K . When you follow either procedure above, the Input Data node is created as shown in Display 2.27B.


Predictive Modeling with SAS Enterprise Miner: Practical Solutions for Business Applications

Display 2.27B °3 chapter^ • J Data Sources - J^J Diagrams -S-fj Ch2_TranVar . Ch2_VarSel RESPONSE_B ^ RESPONSE J j Model Packages 3 Ai Users

• Value

Property

r-Variables rbectslons [Output Type •-Rote "Rerun

I 1 1: 1' 1

-

ds

Node ID Imported Data Exported Data

NN_RESP_DATA • c

View Raw No

i-Data Selection '•New Table

Data Source

HData Source [-Table r Library r.Descriptjpri ;• Number of Observations

NN_RESP_DATA NtJ RESP DATA FHEBOOl

^

29904 18 No

Number of Columns -External Data -Sampling Rate -Segment -Scope

Local

i-Time of Creation i-Run id j-Last Error j-Last Status : [-Needs Updating

12/5/06 8:25 AM

Yes

Display 2.28 <| Select Data Source Name Cr>2 VarSel l')! .'.l l j . ' H TESTIMPUTE

ID CH VARSEL NMRE5PDATA — TESTIMPUTE

Table

Variables

BINARYTARGET NN_RE5P_DATA TESTIMPUTE

OK

Cancel

Help


C 'hapter 2: Getting Started with Predictive Modeling

37

2.6.2 Tools f o r Initial Data Exploration To explore the data, you can use the Enterprise Miner nodes MultiPlot and StatExplore. These nodes enable you to examine the distributions o f the input variables and their relationships with the target variable. In order to use these nodes you must create a process flow diagram, as shown in Display 2.29. To include MultiPlot and StatExplore nodes in the process flow diagram, click on the Explore tab in the node groups (see Display 2.5). The exploration tools are now visible on the toolbar. When you place your mouse pointer over a specific exploration tool, the tool's name becomes visible. Drag the StatExplore and MultiPlot tools onto the Diagram Workspace and connect the nodes as shown in Display 2.29. Display 2.29

UlllPIOl

2.6.2.1 S t a t E x p l o r e N o d e The StatExplore node can be used to find out which input variables are closely related to the target. When you run the StatExplore node and open the Results window, you w i l l see a graph window and an output window. The graph window, shown in Display 2.30A, is called C h i - S q u a r c Plot. This plot shows the statistic known as Cramer's V on the vertical axis and the inputs on the horizontal axis. A Chi-Square plot is shown for categorical targets. Display 2.30A | i Chi Si uore Plot Target = tesp 0.20>

0.15¬

11; | 0.10¬ 2 t^ Ordered Inputj » i r i . n i ' i V • 1-1(114* K1TOE

1*

u.uu-

,— 2

3

1 4

5

— r — 7 6

r ~ i i

i i — ~ i i — i i — h — i , — 1 1 — ,

8 9 10 Ordered Inputs

11

12

13

14

15

IB


38

Predictive Modeling with SAS Enterprise Miner: Practical Solutions for Business Applications

When you place the mouse pointer over a bar, you can see the variable name and the Cramer's V value for that variable. In this example, the top variable is RESTYPE. This variable indicates the customer's residence type (homeowner or renter). Cramer's V for RESTYPE is 0.165347. In general, Cramer's V measures the strength o f relationship between two categorical variables. A detailed explanation o f Cramer's V is given in the appendix to this chapter. Cramer's V is automatically generated for all categorical inputs when the target is also categorical. T o calculate Cramer's V for continuous inputs such as A G E , you must convert the variable from continuous to categorical by creating intervals, or bins, with each bin representing a category. To accomplish this, set the Interval Variables property o f the StatExplore node to Yes, and specify the Number of Bins property to the desired number o f bins, or leave it at its default value o f 5, as shown in Display 2.30B. To get a Chi-Square plot, make sure that the C h i Square property o f the StatExplore node is set to Yes. Display 2.30B Property Node ID Imported Data Exported Data Variables Use Seamen! Variables

Value Stal

No

r Hide Rejected Variables Yes '•Number of Selected Variables 1000 : l-Chl-Square r Interval Variables 1 "[Number.of Bins . ,

'

Yes Yes i5

: j-Correlations [•Pearson Correlations "Spearman Correlations

Yes Yes

[•Time of Creation : [ Run Id [• Last Error [•Last Status [Needs Updating [ Needs to Run [Time of Last Run j [ Run Duration "Grid Host

12/2/06 2:1 GPM

Eip§!9HHHHHH

fes Yes

_ _

For example, the interval-scaled variable A G E is grouped into five bins, which are 18-32.4, 32.4-46.8, 46.8-61.2, 61.2-75.6, and 75.6-90. In the appendix to this chapter, a step-by-step illustration o f the calculation o f the Chi-Square statistic and Cramer's V is shown for the variable AGE. The Chi-Square value shows the strength o f relationship between the target and the binned variable. More specifically, it shows i f the distribution o f the target levels differs significantly from bin to bin among the bins constructed from the interval variable. From the output sub-window in the Results window o f the StatExplore node, you can see the modal value o f each input for each target level. For the input RESTYPE, the modal values are shown in Output 2 . 1 .


Chapter 2: Getting Started with Predictive Modeling 39

Output 2.1 Variable-RESTYPE Target Value

Target

Huncat

Bode Pet

KHlas

_ OVERALL^ reap reap

HOHE HOHE RENTER

54.73 60.19 52.05

Mode2 RENTER REKTER HOHE

40.42 35.11 42.77

This output reports the modal values o f the inputs by the target levels. In this example, the target has two levels: 0 and 1. The columns labeled Mode Pet and Mode2Pct exhibit the first modal value and the second modal value, respectively. The first row o f the output is labeled _ O V E R A L L _ . The _ O V E R A L L _ row values for Mode and Mode Pet indicate that the most predominant category in the sample is homeowners, indicated by H O M E . The second row indicates the modal values for non-responders, and the third row shows the modal values for the responders. The first modal value for the responders is RENTER, suggesting that the renters in general are more likely to respond than home owners in this marketing campaign. These numbers can be verified by running PROC FREQ from the Program Editor, as shown in Display 2.30C. Display 2.30C f ÂŁ 3 Program Editor -

l

E

p r o c ÂŁreq data=TheBook.NN_RE5P_DATA;

2

c a b l e RESTYPE*RESP/nopeccent n o r o u m i s s i n g ;

run ;

3

The results o f PROC FREQ are shown in Output 2.2. Output 2.2 ga output l The FREQ Procedure T a b l e o r RESTYPE b y r e s p RESTYPE

resp

Frequency I Col P e t I

0|

II

Total

CONDO

I I

380 I 1.65 I

186 I 1.98 I

566

COOP

I I

535 I 2.85 I

300 I 3.20 I

885

HOHE

| I

12354 I 60.19 I

4011 I 42.77 I

16365

RENTER

I I

7206 I 35.11 I

4862 I 52.05 1

12088

Total

20525

9379

29904


40

Predictive Modeling with SAS Enterprise Miner: Practical Solutions for Business Applications

2.6.2.2 MultiPlot Node The MultiPlot node can be used for visualizing the data. Y o u can examine the distributions o f the variables and relationships among variables. You must make sure that the Property Sheet is set to Advanced. The property sheet can be set to Advanced by clicking View->Property Sheet-> Advanced. To explore the variables, click on 2

located to the right o f the Variables property. This opens a

window with all the variables, as shown in Display 2.31.

iVariables-Plot Use

Role

Defautt Default Default Default Default Default Default MILEAGE Default MOB Default Default MRTGI MS Default MUMTR Default RESTYPE Default RES_STA Default cuscode Default resp Yes

Input Input Input Input Input Input Input Input Input Input Input Input input Input Rejected Target

Lf

ame

DELINO DEPC EMPâ&#x20AC;&#x17E;STA GENDER HEQ INCOME MFDU

Level Nominal Binary Nominal Binary Nominal Nominal Binaiy Interval Binaiy Nominal Nominal Nominal Nominal Brtary Nominal Binary

Type

Label

Order

Form

h t b b N

N

E N b P b c c c

N

Explore..,

OK

Cancel

Help

To examine the histograms, select the variable o f interest and click on Explore. Display 2.32 shows the histograms generated for the variables RESTYPE and RESP.


Chapter 2: Getting Started with Predictive Modeling

41

Display 2.32 explore - CMWS.Ids_DATA Ffc

View

Actions Window

•JO|4 _JDJ2< Property

is&aMrlrBi Sample Method Fetch Size Random Seed Fetched Rows

Top

Dei&u* 10000

Apply

HOME

Plot...

RENTER

CONDO

RESTYPE _ •

.ids_DATA RESTYPE HOME

x £3

8000-1

1

HOME HOME HOME

HOME HOME

s u. 2000-

RENTER

HOME

RENTER'

RESP

To examine the relationship between a categorical input such as RESTYPE and the target variable RESP, click the Plot button located at the bottom o f the Sample Properties pane in the Explore window shown in Display 2.32. Then the Select a C h a r t Type window appears, as shown in Display 2.33. Display 2.33 Select a Chart Type

§9 Area Bar Histogram C P *

Bar c h a r t provides s t a t i s t i c a l r e l a t i o n s between c a t e g o r i c a l values. Bar can be subdivided and/or grouped u s i n g subgroup v a r i a b l e and/or group v a r i a b l e r e s p e c t i v e l y .

I * Scalier ! 3 Tables sk Matrix Sr| Lattice Parallel Axis h 3D Charts

FWsh

Select the bar chart and click on Next>. The Select C h a r t Roles window w i l l open, as shown in Display 2.34.


Predictive Modeling with SAS Enterprise Miner: Practical Solutions for Business Applications

Display 2.34 Select Chart Roles

Response Category

RESTYPE

Type Numeric Character

I Description RESP RESTYPE

Response statistic: Cared

«Back

By clicking in the box under Role, you get a list o f roles that can be assigned to the variables listed in the first column o f the window shown in Display 2.34. In this example, I set the role o f the variables RESP and RESTYPE to Response and Category, respectively. I set the Response statistic to Mean, clicked Next> four times, and then clicked on Finish. The resulting chart is shown in Display 2.35. Display 2.35 j • - Explore - EMWS.lds_.DATA FJe Edit

View

Actions Window

| | ' Graph3 0.50-

0.40-

S0.30

£o.20 RESTYPE 0.10-

RESP(Hean)

RENTER

-0.414891

0.00

HOME

RENTER

CONDO

COOP

RESTYPE

By passing the mouse pointer over the bars, you can see the response rate for a given level o f the categorical input. From Display 2.35, you can see that the response rate is 4 1 % among renters. Y o u can also examine the relationship o f the inputs to the target variables by setting the Type of C h a r t property o f the MultiPlot node to Bar Charts, and then running the MultiPlot node. By opening the results, you can see charts such as the one shown in Display 2.36.


Chapter 2: Getting Started with Predictive Modeling

43

Display 2.36

R E S T Y P E by resp

F R E Q U E N C Y

56000 16000 14000 13X0 1C000 EOO0

>S36E

eoco

4000 2000 Ol

555

CONDO

CCOP

HOME

RENTER

RESTYPE resp

0

1

The examples given above demonstrate some o f the capabilities o f MultiPlot and StatExplore. For more details, see the Enterprise Miner help.

2.6.3

Impute Node

The Impute node is used for imputing missing values o f inputs. This node icon, along with other icons o f the Modify group, appear on the toolbar when you select the Modify tab o f the Enterprise Miner window (see Display 2.5). Display 2.37 shows the Impute node selected.


44

Predictive Modeling with SAS Enterprise Miner: Practical Solutions for Business Applications

Display 2.37 j enterprise Miner - chapter? He

Lrk

View

Actions

Opt it™

Window

Help

si«sz •

n

*

S « r , * Explore I M o d t y | MoOW j A i i e a i j L U t y | rSiEsinfuiE

M Diagram;

O S

1

DSii S3 &

±1

^RESPONSE • J r*)Sel P s c t a j e ; >

Sl3tE)plore

J . J Users

N N R E S P DAT.

Data Partition

' • D e i s m Ta/jet Method: rw i'Delouri h p u M e t h r x I K t e o i w i | D e l a u l TBget Methoflone [•AEWTimg

90

[• AHUBER Turlno

'l 5

r A W A V E Turwg

6?B3ir»307Z

;

StXffirjPlOficrlicn

WO

E } Leal a r e

Default Input Method

Specifies the imputation method for interval input variables Didisjini RESf 0 ( HE Onyi t - J

h i \ C o m e d W t o SAEMam -Locngal W a k i p a c e Server

The Impute node has a number o f techniques for imputation. To select a particular technique, you must first select the Impute node on the Diagram Workspace and set the Default Input Method property to the desired technique. For interval variables, the options available are Mean, Median, Mid-Range, Distribution, Tree, Tree Surrogate, M i d - M i n i m u m space, Tukey's Biweight, Huber, Andrew's wave, Default constant, and None. For class variables, the options are Count, Default constant value, Distribution, Tree, Tree surrogate, and None.

2.6.4

Data Partition Node

In model building it is necessary to partition the sample into Training, Validation, and Test sub-samples. The training data is used for developing the model using tools such as Regression, Decision Tree, and Neural Network. During the training process, these tools generate a number o f models. The validation data set is used to evaluate these models, and then to select the best one. The process o f selecting the best model is often referred to as fine tuning. The test data set is used for an independent assessment o f the final model. The Data Partition node is shown in Display 2.38.


Chapter 2: Getting Started with Predictive Modeling 4 5

Display 2.38 j Enterprise Miner - chapter2 He

Edt

V*w Acuoru Conors

Wndow Hetn

S n S â&#x20AC;˘ â&#x20AC;˘ [_3 ill Sarrptol EiMnej t-toaty] Moddj Assess| UUtyj

In the Data Partition node, you can specify the method o f partitioning by setting the Partitioning Method property to one o f four values: Default, Simple random. Cluster and Stratified. In the case o f a binaiy target such as response, the stratified sampling method results in uniform proportions o f responders in each o f the partitioned data sets. Hence, I set the Partitioning Method property to Stratified, which is the default for binaiy targets. The default proportion o f records allocated to these three data sets are 40%. 30%. and 30%, respectively. These proportions can be changed by resetting the Training. Validation, and Test properties under the Data Set Percentage property. In order to reset the properties, you should first drag the Data Partition node into the Diagram Workspace, and select it, so that the Property Panel w i l l show the properties o f the Data Partition node as shown in Display 2.38.

2.6.5 Filter Node The Filter node can be used for eliminating observations with extreme values (outliers) in the variables. You should not use this node routinely to eliminate outliers. While it may be reasonable to eliminate some outliers for very large data sets for predictive models, the outliers often have interesting information that leads to insights about the data and customer behavior. Before using this node, you should first find out the source o f any extreme value. I f the extreme value is due to an error, the error should be corrected. I f there is no error, you can truncate the value so that the extreme value does not have an undue influence on the model. Display 2.39 shows the flow diagram with the Filter node.


46

Predictive Modeling with SAS Enterprise Miner: Practical Solutions for Business Applications

Display 2.39

NN RESP DATA

Data Partition

Filter

In the process flow shown in Display 2.39. the Filter node follows the Data Partition node. Alternatively, you can use the Filter node before the Data Partition node. To use the Filter node, select it. and set the Default Filtering Method property for Interval variables to one o f the values listed there as shown in Display 2.40A. Display 2.40A Mean Absolute Deviation (MAD) User-Specified Limits Metadata Limits

i

Extreme Percentiles Modal Center Standard Deviations from the Mean None

For class variables, the Default Filtering Method property can be set to one o f the values as shown in Display 2.40B. Display 2.40B iRare Values (Count) Rare Values (Percentage) JNone

Display 2.40C shows the properties panel o f the Filter node.


Chapter 2: Getting Started with Predictive Modeling

Display 2.40C - 1 — ' Property Nods ID Imported Dala Exported Data Exncrt Table TaGles to Filter Create score code

Vatue Filler TJ

Filtered All Dala Sets

N

Yes

H

-C!ass Variables ;• Default Filtering Method Rare Values (Percentage) Yes j-Keep Missing Values Yes [-NormalizedValues F Minimum Frequency Cutoff 1 [ Minimum CutoffforPereentacO 01 Maximum Number of Levels '25 ;

EL

interval Variables Default Filtering Method Exlreme Percentiles Keep Missing Values "rfes •Cutofffor MAD •CutofTPercentiles for ExtiemtO 5 -CutoffforModal Center 9.0 -Cutoff for standard Deviation 3.0

[ Time of Creation I-Run Id iLast Error [-Last Status j-Needs Updating [•Needs to Run

12(6706 4:29 PM c8E3aac6-93a6-48aa-b58tH Complete No No

I set the Tables to Filter property to All Data Sets so that outliers are filtered in all three data sets—training,validation, and t e s t — a n dthen ran the Filter node. The Results window displays the number o f observations eliminated due to outliers o f the variables. Output 2.3 A (from the output o f the Results window) shows the number o f records excluded from the T r a i n , Validate, and Test data sets due to outliers. Output 2.3A Numben o f O b s e r v a t i o n s Table TRAIN VALIDATE TEST

Data

Included

11960 8972 8972

11557 8658 8673

Excluded 403 314 299

Outputs 2.3 to 2.5 show summary statistics for original and filtered data for the variable A G E .


48

Predictive Modeling with SAS Enterprise Miner: Practical Solutions for Business Applications

Output 2.3 j

D a t a Role=TRAIN Variable=AGE Statiscics Number H i s s i n g Minimum Maximum He an Std D e v i a t i o n Skeuness Kurtosis

Original

Filtered

j

0.0000 18.0000 90.0000 49.6672 15.7092 0.1867 -0.70S8

0.0000 22.0000 88.0000 49.6075 15.3838 0.1598 -0.7781

[

r

!

\

.„.,.,.„—,—

Output 2.4 Data Role"VALIDATE Variable =AGE

:

Statistics Number Hissing Minimum Maximum He an Std Deviation Skeuness Kurtosis

Original

Filtered

s

0.0000 18.0000 90.0000 49.1291 15.4426 0.2047 -0.6851

0.0000

1

22.0000 88.0000 49.0687 15.1197 0.1799 -0.7676

j [ | | \ 1

Output 2.5 t

Data Role=TEST Variable=AGE Statistics Number H i s s i n g Minimum Haximum He an Std D e v i a t i o n Sketmess Kurtosis

2.6.6

Original

Filtered

0.0000 18.0000 90.0000 48.9826 15.4225 0.2069 -0.7138

0.0000 22.0000 88.0000 49.0687 15.1197 0.1799 -0.7676

\

Variable Selection Node

The Variable Selection node can be used to make a preliminary selection of the variables to be included in predictive modeling. There are a number of alternative methods with various options for selecting variables. The methods of variable selection depend on the measurement scales of the inputs and the targets. These options are discussed in detail in Chapter 3. In this chapter, I present a brief introduction to the techniques used by the Variable Selection node for different types of targets, and I show how to set the properties of the Variable Selection node for choosing an appropriate technique. There are two basic techniques used by the Variable Selection node. They are the R-Square selection method and the Chi-Square selection method. Both of these techniques select variables


C 'hapter 2: Getting Started with Predictive Modeling

49

based on the strength o f their relationship with the target variable. For interval targets, only the R-Square selection method is available. For binaiy targets, both the R-Square and Chi-Square selection methods are available.

2.6.6.1 R - S q u a r e S e l e c t i o n Method To use the R-Square selection method, you must set the Target Model property to R-Square. as shown in the Display 2.41. Display 2.41 Property Node ID Imported Data Variables Max Class Level Max Missing Percentage Target Model Hide Rejected Variables

Value Varsel Mi

100 50 R-Square Default

Reject Unused Variables ! [-Number of Bins I [-Maximum Pass Number i -Minimum Chi-Square

Chi-Square R and Chi Square None

PSSESSSSS239I

[-Maximum Variable Number | [-Minimum R-Square : [-Stop R-Square j [-Use AOV15 Variables i [-Use Group Variables -Use Interactions j jSPDS ! [-Last Error j [-Last Status i [-Needs Updating j ['Needs to Run j [-Time of Last Run -Run Duration ;

0.0050 1 0E-5 Yes r-es

to Yes

Complete Mo No I0J3/D5 10:39 PM

OHr.QMin.11.31 Sec.

When the R-Square selection method is used, the variable selection is done in two steps. In Step 1, the Variable Selection node computes an R-Square value (the squared correlation coefficient) based on the relationship between the target and each input variable, and then assigns the Rejected role to those variables that have a value less than the minimum R-Square. The default minimum R-Square cut-off is set to 0.005. You can change the value o f the minimum RSquare by setting the Minimum R-Square property to a value other than the default. The R-Square value (or the squared correlation coefficient) is the proportion o f variation in the target variable explained by a single input variable, ignoring the effect o f other input variables. In Step 2, the Variable Selection node performs a forward stepwise regression to evaluate the variables chosen in the first step. Those variables that have a stepwise R-Square improvement less than the cut-off criterion have the role o f rejected. The default cut-off for R-Square improvement is set to 0.0005. This can be changed by setting the Stop R-Square property to a different value.


50

Predictive Modeling with SAS Enterprise Miner: Practical Solutions for Business Applications

For interval variables, the R-Square value used in Step 1 is calculated directly by means of a linear regression of the target variable on the interval variable, assessing only the linear relation between the interval variable and the target. To detect nonlinear relations, the Variable Selection node creates binned variables from each interval variable. The binned variables are called A O V 1 6 variables. Each A O V 1 6 variable has a maximum of 16 intervals of equal width. The A O V 1 6 variable is treated as a class variable. A one-way analysis of variance is performed to calculate the R-Square between an A O V 1 6 variable and the target. These A O V 1 6 variables are included in Step 1 and Step 2 of the selection process. Some of the interval variables may be selected both in their original form and in their binned ( A O V 1 6 ) form. If you set the Use A O V 1 6 Variables property to Y e s , then only A O V 1 6 variables appear in the Variable Selection pane of the Results window, and these will be passed on to the next node, with their Role set to Input. For variable selection, the number of categories of a nominal categorical variable can be reduced by combining categories that have similar distribution of the target levels. To use this option, set the Use G r o u p Variables property to Yes.

2.6.6.2 Chi-Square Selection Method When you select the Chi-Square selection method, the Variable Selection node creates a tree based on Chi-Square maximization. The Variable Selection node first bins interval variables and then uses the binned variable rather than the original inputs in building the tree. The default number of bins is 50, but the number can be changed by setting the Number of Bins property to a different value. Any split with a Chi-Square below the specified threshold is rejected. The default value for the Chi-Square threshold is 3.84. This number can be changed by setting the Minimum Chi-Square property to the desired level. The inputs that give the best splits are included in the final tree, and passed to the next node with the role of Input. A l l other inputs are given the role of Rejected.

2.6.6.3 Variable Selection Node: An Example with R-Square Selection A s a means of illustrating how our results depend on the settings of the properties of the Variable Selection node, I experiment with a small set of variables and show you the results below. Display 2.42 shows the process flow with the Variable Selection node.

Display 2.42

Ch2_VarSel

Data Partition

Variable Selection

Regression

Display 2.43 shows the Variables window of the Variable Selection node. The data set is partitioned such that 50% is allocated to training, 30% to validation, and 20% to test.


Chapter 2: Getting Started with Predictive Modeling

51

Display 2.43 ,-',! Variables - Varsel

Name •AR1N VAR20 VAR3N VAR4C VAR50 VAR6C VAR7C V-AR80 credscore resp

Use Default Default Default Default Default Default Default Default Default Yes

Role

Level

input input Input Rejected Input Rejected Rejected Input Input Target

Interval Nominal Nominal Nominal Nominal Nominal Nominal Nominal Interval Binary

Order

Type

Label

N N N C

Format

USAGE WEALTH PRC CLUSTER Credit score

fcqi've.

OK

Cancel

Hefc

Since the target variable is binary in this example, any o f the above options shown in Display 2.44 could be used. I chose the R-Square selection method by setting the Target Model property to R-Squarc. Display 2.44 1

Value

Jfroperty ^ Target Model Hide Rejected Variables

R-Square

Reject Unused Variables •dJiiLUUJ,!. I S i

R-Square

Default M Chi-Square

[Number of Bins [-Maximum Pass Number ill*-.

R and Chi Square

\

None

Next, I select the threshold values for the Minimum R-Square and Stop R-Square properties. The Minimum R-Square property is used in Step 1, as described in Section 2.6.6.1. The Stop R-Square property is used in Step 2, which is also described in Section 2.6.6.1. In this example, the Minimum R-Square property is set to the default value, and the Stop R-Square property to 0.00001. The Stop R-Square property is deliberately set at a very low level so that most o f the variables selected in the first step w i l l also be selected in the second step. In other words, I effectively by-passed the second step so that all or most o f the variables selected in Step 1 are passed to the Tree node or Regression node that follows. I did this because both the Decision Tree and Regression nodes make their own selection o f variables, and I wanted the second step o f variable selection to be done in these nodes rather than in the Variable Selection node. Display 2.45 [•Maximum Variable NumfcCOOO [-Minimum R-Square 0.0050 [-Stop R-Square [-Use AOV16 Variables [ Use Group Variables -Use Interactions SPDS

1 .OE-5 Yes Yes No Yes


Predictive Modeling with SAS Enterprise Miner: Practical Solutions for Business Applications

From Display 2.45, you can see that 1 set the Use A O V 1 6 Variables property to Yes. I f I chose No. the Variable Selection node (i.e., the underlying PROC D M I N E ) would still create the A O V 1 6 variables and include them in Step 1, but not in Step 2. In addition, i f you select No, these variables w i l l not be passed to the next node. When you set the Use Group Variables property to Yes, the categories o f the class variables are combined, and new variables with the prefix o f G_ are created. (This is explained in more detail in the next chapter.) With the settings shown in Display 2.45,1 ran the Variable Selection node, and after the run was completed I opened the Results window by selecting the Variable Selection node, right-clicking on it. and then clicking on Results. Display 2.46 shows the Results window. Display 2.46

MM

The bar chart in the top left quadrant o f Display 2.46 shows the variables selected in Step 2. The bottom left quadrant shows the variables selected in Step I . The output shown in the top right quadrant shows the variables selected in Step 2. The output pane in the bottom right quadrant o f the Results window shows the variables selected in Step 1 and Step 2. By scrolling down the output pane, you first see the variables selected in Step 1. These also appear in Output 2.6.


Chapter 2: Getting Started with Predictive Modeling 53

Output 2.6 The DHINE Procedure R-Squares f o r T a r g e t V a r i a b l e : resp Effect

DF

R-Square 0.010908 0.010724 0.009589 0.009517 0.009424 0.009410

C l a s s : VAR80

17 5 8 3 8 4 6

Group: A0V16: Var: A0V16: Var:

2 14 1 3 1

Class: Group: Class: Group: Class: Group:

VAR3N VAR3N VAR50 VAR50 VAR20 VAR20 VAR80 VAR1N VAR1N credscore credscore

0.008450 0.008392 0.006479 0.003508 0.000289 0.000162

R2 < HIHR2 R2 < HIKR2 R2 < HINR2

;

Scrolling down further in the output pane, you can see the variables selected in Step 2. These are shown in Output 2.7.

Output 2.7 TTÂťe DHIHE Procedure Effects Chosen for Target: resp

Effect

DF

R-Square

F Value

p-Value

Sun of Squares

Group: VAR3N A0V16: VAR1N Group: VARS0 Group: VAR20 Group: VAR80

5 14 3 4 2

0.010724 0.003912 0.001891 0.000673 0.000567

19.369933 2.529776 5.715847 1.525318 2.574017

<.0001 0.0013 0.0007 0.1918 0.0763

8.628783 3.14790S 1.521686 0.541304 0.456572

Error He an Square 0.089095 0.088882 0.088741 0.088720 0.088689 i

The variables shown in Output 2.7 were selected in both Step 1 and Step 2. The variables selected by the Variable Selection node are given the role of Input and passed to the next node. The Variable Selection area in the top right quadrant of the Results window (shown in Display 2.46) shows the variables with their assigned roles. The next node in the process flow is the Regression node. By opening the Variables window of the Regression node, you can see the variables passed to it by the Variable Selection node. These variables are shown in Display 2.47.


54

Predictive Modeling with SAS Enterprise Miner: Practical Solutions for Business Applications

Display 2.47 w

Variables - Reg

Name A0V16_VAR1N G_VAR"5O G~VAR3N G_VAR50 G_VARBO

resp

j

Use Default Default Default Default Default Yes

*1 Report No No No No No No

! Role .Input â&#x20AC;˘Input

Input Input

jnput, Target,

Level Ordinal Nominal Nominal Nominal Nominal Binary

I Type N

Order

Lab

IN

1 Explore...

OK

Cancel

Help

Display 2.48 (taken from the output area o f the Results window) shows that V A R I N is rejected in Step I , but A 0 V 1 6 _ V A R 1 N is selected in both Step 1 and Step 2, indicting a nonlinear relationship between V A R 1 N and the target. The definitions o f the A O V 1 6 variables and the grouped class variables (G variables) are included in the SAS code generated by the Variable Selection node. This code can be accessed by selecting the Variable Selection node, right-clicking on it, and clicking on Results-^ View->Scoring->SAS Code. Display 2.48 shows a segment o f the SAS code.


Chapter 2: Getting Started with Predictive Modeling

Display 2.48 1 2 3

*** Begin Scoci n,g Code from PROC DMIHE rwifirii J!HtJtr«tmUH«UtJtHti.

4 5 6 7

l e n g t h JJARN . S 4; l a b e l _WARH_ = "WdiTiing's"; JJARH_ = ' '

8 9

10 11

12 13 j 14 • 1 5

I

1 6

17 18 19 20 21 22 23 24 25 | 26 27 . 28 | 29 30 31 j 32 33

34 35 36 37 j 38 39

i f (VAR1N = else i f (VAR1N <= else i f (VAR1N <= else i f (VAR1N <= el3e i f (VAR1N <= else i f (VAR1H <= else i f (VAR1N <= else i f (VAR1N <= el3e i f (VAR1N <= else i f (VAR1N <= else i f (VAR1N <= else i f [VAR1H <= el3e| i f (VAR1N <= else i f {VAR1N <= else i f (VAR1N <= else i f (VAR1N <=

) t h e n A0V16_VAR1N = 3; 3 . 1)125) t h e n A0V16_VAR1M =

1;

10 .625) t h e n A0V16_VAR1N • 2 ; 15 .4315] t h e n A0V16_VAR1H = J; 20 .25) t h e n A0V16_VAR1N = 4; 25 .0625) t h e n A0V16_VAR1N - 5; 29 .875) t h e n A0V16_VAR1H - r>; 34 6875) t h e n A0V16_VAR1N = i; 39 5) t h e n A0V16_VAR1N = 1 44 3125) t h e n A0V16_VAR1N = 9; 49 .125) t h e n A0V16_VAR1M = 10; 53 9375) t h e n A0V16_VAR1N • 1 1 ; 58 75) t h e n A0V16_VAR1N = 12; 63 5625) t h e n A0V16_VAR1H = 13;

eg

375) t h e n A0V16_VAR1M

« 14;

73 1875) t h e n A0V16_VAR1K = 15;


'6VZ ^B|dsiQ ui UMoqs si ssjqBUBA aq; j o aouBjjodiui 9Ai;Bj9i a q x 'saiqeuBA p s j o a p s a q i a i u o o s q gqj ui pasn sajqeuBA a q j puB

'99.11jo

adAj Q I V H D

B

sjonnsuoo j s u p / j a s u d i a j u g

o j Ajjsdojd p p o j \ i ; 3 3 J B X s q j J9S 'uouaujo s j B n b s - i q o

uou,oa|as ajenbs-mo

MJIM

a j E n b g - i q ^

Suisn u o ; j 0 9 | 9 S 9 | q B U B A u u q r j s d o j ,

9 q j

a|diuexg uv :apoN uojioaias aiqeueA

P'9'9'Z

rptra ' = NCOTA~9 ( J s

NCOTA~3 t J ( ,

= N£HVA~9

i

H„)iraittt H£HVA~£I ( i i *s = MeWA~3 ( . = KCHVA~9 ( ,

1 j

N£OTA~9 ( i I ! I \

1 I j

iz

= N£HVA~9

N£tfVA~9 'T N£tfVA~3 _ = H£HVA 3 iX ?T = N £ H V A ~ 9 fZ s N £ H Y A ~ 3 _ = K£HVA 3 N£HVA~9 i

z

(, ( i ( , ( « (, ( * ( i ( ,,

9„)uaqn

Z„)xxavia.

= N£HVA~£) ( i = N£HVA~9

( J

>

i

.'(ZTIMOff k o a l a s ( ZTHttOlT )<ZDQKUOBtt NCHVA )^nd = 2THH0N

i

•'2T J ZTHb!0N~ usfiusT

(penuiiuoo)

suotjaatjddy ssdutsnq uof suotptjog JDOIJODUJ

uamy^ dsududiug SVS

M

%yz

AeidsjQ

Ht! Suippojv

d\tjDipduj

9c,


Chapter 2: Getting Started with Predictive Modeling

57

Display 2.49 Variable Importance

1.0 -

0.8u c n

o 0.6 Q.

E 0.4rr

0.2-

0.0

I

VAR BO

VARI1N

VAR 11

/AR50

VAR20

Variable

Display 2.50 shows the variables selected by the Chi-Square selection process. Display 2.50 if IB Variable Selection Name credscore /AR1N V-AR20 V-AR3N •AR4C VAR50 •AR6C VAR7C •AREO

Role Rejected Input Input Input Rejected Input Rejected Rejected Input

Lev.:-! Interval Interval Nominal Nominal Nominal Nominal Nominal Nominal Nominal

Type N N N N C N C C N

H

i

Label

Credit score

USAGE WEALTH PR. CLUSTER

Comment Var s el SmeJI...

2.6.6.5 S a v i n g the S A S C o d e G e n e r a t e d by the V a r i a b l e S e l e c t i o n N o d e I f the Target Model property is set to R-Square. the Use A O V 1 6 Variables property is set to Yes, and the Use G r o u p Variables property is set to Yes. then the Variable Selection node w i l l generate SAS code with definitions o f the A O V l 6 and Group variables. When you right-click on the Variable Selection node and then click on Results, the Results window opens. I f you click View S A S Results ^Flow Code in the Results window you can view the SAS code. A partial listing o f this code is shown in Display 2.51. -


58 Predictive Modeling with SAS Enterprise Miner: Practical Solutions for Business Applications

Display 2.51 \ *** Begin Scoeing Code tvou

PROC DHIHE

; length _UARN_ « 4; l a b e l _OARN_ «* "Warnings"; ; UARN = ' '; i f (VARIH » .) then A0V16_VAR1N - 3; ' el3e i f (VAR1N <- 5.8125) then A0V16_VAR1H »1; ;;else i f (VAR1N <= 10.625) then A0V16_VAR1H »2; ;;else | ; i f (VARIH <= 15.4375) then A0V16_VAR1N = 3; : else ! \ i f (VARIH <= 20.25) then A0V16_VAR1H »4; i else [ i f (VAR1N <- 25.0625) then A0V16_VAR1H - 5;

•Seise i f (VARIH <«29.875) then A0V16_VAR1H »6; ; else | i f (VARIH <= 34.6875) then A0V16_VAR1N «• 7; j else) | i f (VARUS <= 39.5) then A0V16_VAR1N = 8; s else [ i f (VARIH <»44.3125) then A0V16_VAR1N = 9; i else j i f (VAR1H <- 49.125) then A0V16_VAR1H - 10; i else i i f (VAR1N <»53.9375) then A0V16_VAR1H " JUL; | else \ i f (VARlil <= 58.75) then A0V16_VAR1H = 12; J else j; i f (YAR1H <= 63.5625) then A0V16_VAR1H = 13; ; else ! i f (VARIH <»68.375) then A0V16_VAR1N = 14; \ else [ i f (VARIH <- 73.1875) then A0V16_VAR1N - 15; e l s e A0V16_VAR1H = 16;

This can be saved by clicking File-^Save as and giving a name to the file where the SAS code is being saved and pointing to the directory where the file is being saved. The code can also be saved by clicking Results View Scoring SAS Code, followed by File->Save as.

2.6.6.6 The Procedures Behind the Variable Selection Node Enterprise Miner uses P R O C D M I N E in variable selection. Prior to running this procedure, it creates a data mining database ( D M D B ) and a data mining database catalog ( D M D B C A T ) using P R O C DMDB. In the Results window of the Variable Selection node, you can see the log by clicking on View on the menu bar and selecting log. Display 2.52 shows the log from P R O C DMDB.


Chapter 2: Getting Started with Predictive Modeling 59

Display 2.52 Run proc dmdb w i t h the s p e c i f i e d maxlevel c r i t e r i o n .

»

*.

*

* EH: DHDBClass Macro ;

*

*.

%macro DHDBClass; VAR20(ASC) VAR3N(ASC) VAR50(ASC) VAR80(ASC) resp(DESC) %mend DHDBClass;

*

* .

* EH: DHDBVar Hacro ;

»

» .

%macro DHDBVar; VARIH credscore *mend DHDBVar;

*

*•

* EH: Create DHDB;

«

*;

libname _ s p d s l i b SPDE "C:\DOfJTJHE~lNsasadm\LOCALS~l\Temp\SAS Temporary Files\_TD4908" L i b r e f _SPDSLIB was s u c c e s s f u l l y assigned a s f o l l o w s : Engine: SPDE P h y s i c a l Name: C:\D0COHE~l\sasadm\L0CALS-l\Temp\SAS Temporary Files\_TD4908\ proc dmdb b a t c h data°EHErsl.Part_TRAIN dmdbcat=H0RK.EH_DHDB maxlevel ° 101 o u t » _ s p d s l i b . EH_DHDB r

c l a s s %DHDBClass; var %DHDBVar; target resp run;

L Display 2.53 shows the

log from PROC DMINE.

Display 2.53 * Varsel: Input Variables Hocno ; %macco INPUTS; VARIH VAR20 VAR3N VAR50 VAR80 CREDSCORE %mend INPUTS; proc dnine data=_spd31ib.EH_DHDB dadbcacHORK.EH_DHDB Binr2»0.G05 aaxtowsOOOO stopr.2°0.00001HOINTER USEGROUPS outesc^aORK._Varsel_0UTESTDHINE j] H0H0HTT0R PSHORT var %INPUTS; target resp; , code £ile-'X:\TheBoo^t\EH_S.2\Nov2006^ChapceE2B\«or!lspoces\ErTHSl\VaE3el\EHr^.0HSC0RE.3a3*; code Cile='X:\TheBoo>t\EH_S.2\Nov2006\Chapt4:E2B\Hoi^pace3\Eira31\Vttr3el\EHPUBLISHSC0RE.3a3" cun;

2.6.7 Transform Variables Node Sections 2.6.7.1 and 2.6.7.2 describe the different transformation methods available in the Enterprise Miner Transform Variables node. Sections 2.6.7.3 to 2.6.7.6 show some examples of how to make the transformations, save the code, and pass the transformed variables to the next node.


Predictive Modeling with SAS Enterprise Miner: Practical Solutions for Business Applications

2.6.7.1 Transformations for Interval Inputs â&#x20AC;˘

Simple Transformations The available simple transformations are Log, Square Root, Inverse, Square, Exponential, and Standardize. They can be applied to any interval-scaled input. These simple transformations can be used irrespective of whether the target is categorical or continuous.

â&#x20AC;˘

Binning Transformations In Enterprise Miner, there are three ways of binning an interval-scaled variable. To use these as default transformations, select the Transform Variables node, and set the value

of the Interval Inputs property to Bucket, Quantile, or Optimal in the Default Methods section. o

Bucket: The Bucket option creates buckets by dividing the input into n equal-sized intervals and grouping the observations into the n buckets. The resulting number of observations in each bucket may differ from bucket to bucket. For example if A G E is divided into the four intervals 0-25,25-50, 50-75, and 75-100 then the number of observations in the interval 0-25 (bin 1) may be 100, the number of observations in the interval 25-50 (bin 2) may be 2000, the number of observations in the interval 50-75 (bin 3) may be 1000, and the number of observations in the interval 75-100 (bin 4) may be 200.

o

Quantile: The Quantile option groups the observations into quantiles (bins) with an equal number of observations in each. If there are 20 quantiles, then each quantile consists of 5% of the observations.

o

Optimal Binning for Relationship to Target: This transformation is available for binary targets only. The input is split into a number of bins, and the splits are placed so as to make the distribution of the target levels (for example, response and nonresponse) in each bin significantly different from the distribution in the other bins. To help you understand optimal binning on the basis of relationship to target, consider two possible ways of binning. The first method involves binning the input variable recursively. Suppose you have an interval input X that takes on values ranging from 0 to 320. A new variable with 64 levels can be created very simply by dividing this range into 64 sub-intervals as follows:

0 < X < 5 , 5<X<10,

,310<^<315, 3 1 5 < X < 3 2 0 .

The values the new variable can take are 5, 10, 15... 315, which can be called splitting values. Y o u can split the data into two parts at each splitting value and determine the Chi-Square value from a contingency table, with columns representing the splitting value of the input X and rows representing the levels of the target variable. Suppose the splitting value tested is 105. At that point, the contingency table will look like the following:

*<105

X>105

n

"12

Target level 1 0

u

W

02


Chapter 2: Getting Started with Predictive Modeling

61

where, n ,n ,n , and n denote the number of records in each cell of the contingency table. The computation of the Chi-Square statistic is detailed in Chapter 4. n

l2

Qi

02

The splitting value that gives the maximum Chi-Square determines the first split of the data into two parts. Then each part is split further into two parts using the same procedure, giving two more splitting values and four partitions. The process continues until no more partitioning is possible. That is, there is no splitting value that gives a Chi-Square above the specified threshold. At the end of the process, the splitting values chosen define the optimal bins. This is illustrated in Display 2.54.

Display 2.54 Optimal binning of an interval input

All records (5.000)

X

2.CCD records

3.000 records

Jf < 25

1,000 records

'â&#x20AC;˘

X i. 25

500 records

2,000 records

1,500 records

From Display 2.54, the optimal bins created for the input X are X < 25, 25 < X < 105, 105 < X < 135 and 135 < X . Four optimal bins are created from the initial 64 bins of equal size. The optimal bins are defined by selecting the best split value at each partition. The second method involves starting with the 64 or so initial bins and collapsing them to a smaller number of bins by combining adjacent bins that have similar distribution of the target levels. Enterprise Miner does not follow the steps outlined above exactly, but Enterprise Miner's method of finding optimal bins is essentially what is described in the second method above. There are other ways of finding the optimal bins. Note that in the Transform Variables node the optimal bins are created for one intervalscaled variable at a time in contrast to the Decision Tree node where the bins are created with many inputs of different measurement scales, interval as well as categorical.

â&#x20AC;˘

Best Power Transformations The Transform Variables node selects the best power transformations from among X,\og(X),sqrt(X), e \ X \ X, and X , where X is the input. There are four criteria of "best" available: u

o

2

4

Maximum Normal: T o find the transformation that maximizes normality, sample quantiles from each of the transformations listed above are compared with the theoretical quantiles of a normal distribution. The transformation that yields quantiles that are closest to the normal distribution is chosen.


62

Predictive Modeling with SAS Enterprise Miner: Practical Solutions for Business Applications

Suppose Y is obtained by applying one of the above transformations to X. For example, the 0.75-sample quantile of the transformed variable Y is that value of Y at or below which 75% of the observations in the data set fall. The 0.75-quantile for a standard normal distribution is 0.6745 given by P{Z < 0.6745) = 0 . 7 5 , where Z is a normal random variable with mean 0 and standard deviation 1. The 0.75-sample quantile for Y is compared with 0.6745, and similarly the other quantiles are compared with the corresponding quantiles of the standard normal distribution. o

Maximum Correlation: This is available only for continuous targets. The transformation that yields the highest linear correlation with the target is chosen.

o

Equalize Spread with Target Levels: This method requires a class target. The method first calculates variance of a given transformed variable within each target class. Then for each transformation it calculates the variances of these variances. It chooses the transformation that yields the smallest variance of the variances.

o

Optimal Maximum Equalize Spread with Target Level: This method requires a class target. It chooses the method that equalizes spread with the target.

2.6.7.2 Transformations of Class Inputs For class inputs, two types of transformations are available. â&#x20AC;˘

Group Rare Levels transformation: This transformation combines the rare levels into a separate group, _ O T H E R _ . To define a rare level, you define a cutoff value using the Cutoff Value property.

â&#x20AC;˘

Dummy Indicators Transformation: This transformation creates a dummy indicator variable (0 or 1) for each level of the class variable.

To choose one of these available transformations, select the Transform Variables node and set the value of the Class Inputs property to the desired transformation. By setting the Class Inputs property to Dummy Indicators Transformation, as mentioned before, you will get a dummy variable for each category of the class input. Y o u can test the significance of these dummy variables using the Regression node. The groups that correspond to the dummy variables, which are not statistically significant, can be combined, or omitted from the regression. I f they are omitted from the regression, it implies their effects are captured in the intercept term.

2.6.7.3 Transformations of Targets In the Transform Variables node of Enterprise Miner 5.2, the target variables can also be transformed by setting the Interval Targets and Class Targets properties. I f you click the Value column in the Interval Targets property or Class Targets property, a list of all available transformations drops down and lets you select the desired property.

2.6.7.4 Selecting Default Methods To select a default method for transforming interval inputs, select the Transform Variables node in the Diagram Workspace and set the Interval Inputs property (under the Default Methods section in the properties panel) to one of the transformations described in 2.6.7.1. The selected method will be applied to all interval variables. Similarly, set the Class Inputs property to one of


Chapter 2: Getting Started with Predictive Modeling

the methods described in 2.6.7.2. Enterprise Miner applies the chosen default method to all inputs o f the same type. 2.6.7.5 O v e r r i d i n g the Default M e t h o d s I f you do not want to use the default method o f transformation for a particular input, you can override it by right-ciicking on 0

located on the right side o f the Variables property o f the

Transform Variables node, as shown in Display 2.55. This opens the Variables window as shown in Display 2.56. Display 2.55 3&Mer2

D ' -

rvedei Com [arisen

w

CM TwVn

CfflFira:r

'tendinis

- H e m r»g«

r-CUss " . : = Norn

TOP

•KeflKid -S:;

Maul

-fisr.dcmSsed

ii a n

Bj -CuWVJlue

05

Ml

VM

ftt Ui»r Ss.-r

UrssrqVau

Md1 * * i w i V*a to M H r w x YM Cffss) I'ilui

11

SjrrMryVsrstlfs

Transformed a n d t a » V

Display 2.56 , Variables - Trans Mame

Number of Sins

Method

Role

Level

I

Log

4 Input

Interval

IN

resp

4Target

Binary

N

4 Input

interval

N

Ordinal

N

«R1N

Default Optimal

VAR 2 0

nverse

VAR3N

Square

4 Input 4|nput

VAR4C

Exponential

4 Rejected

Nominal

VAR 5 0

Centering

4 inpul

Nominal

4 Rejected

Nominal

C

4 Rejected

Nominal

Ic

4 input

Nominal

N

VAR6C VAR7C

Standardize Riirlf<>t

WVR80

Quanlile

DUU.IV C I

• • •'

Order

Type

credscore

_

_

j<

Interval

b V

I OK

Cancel

Help

63


Predictive Modeling with SAS Enterprise Miner: Practical Solutions for Business Applications

In the Variables window, click on the Method column for the variable V A R 1 N , and select Optimal. Similarly, select the log transformation for the variable CREDSCORE (Credit score). After these changes are made, the resulting Variables window is shown in Display 2.63, which also indicates that Optimal binning is used for the variable V A R 1 N and log is used for the variable CREDSCORE. Transformations are not performed for other variables since the default methods are set to N O N E in the properties panel (see Display 2.57). Display 2.57 ,V,|Variables - Trans Name

Method

VAR1N VAR 20 ]/AR3N </AR4C VAR 50 VAR6C VAR7C VARSO credscote resp

Number ol Bins

Role 4 Input 4 Input 4jlnput 4 Rejected 4 Inpul 4 Rejected 4 Rejected 4 input 4 Input 4iTarqet

Optimal

Default Default Default Default Default Defaull Default

Low

Default

<l

Level Interval Ordinal interval Nominal Nominal Nominal Nominal Nominal Interval Binary

Type

Order

IN |N N

c

N C C N N N

I

i (

1 Explore.

OK

Cancel

Help

2.6.7.6 S a v i n g the S A S C o d e G e n e r a t e d by t h e T r a n s f o r m V a r i a b l e s N o d e Right-click on the Transform Variables node and run it. Then click on Results. In the Results window, click View->SAS Results-^Flow Code to see the SAS code, as shown in Display 2.58. Display 2.58 ||

Flow Code 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

* Computed Code; *

* TPJJJSFOPH: c r e d s c o r e , l o y ( c t e d s c o r e + 1 ) ; label

LOG_credscore =

'Transforced:

Credit

score';

i f c r e d s c o r e + 1 > 0 t h e n L0G_cr;edscoi:e = l o g ( c r e d s c o r e e l 3 e LOG_credscoce = .; * TPATISFORH: VARIH , Optimal B i n r i i i i g i - l ) ; l a b e l 0PT_VAR1H = 'Transformed l e n g t h 0PT_VAR1N 5 2 1 ; i f (VARIH < 13. â&#x20AC;˘>) t h e n QPT_VAR1H = "01:low

-13. S,

VMW;

mSSING";

else i f (VARIH >= 13.1) t h e n 0PT_VAR1H = " 0 2 : 1 3 . 5 - h i $ h " ;

This code can be saved by clicking FiIe->Save as. and typing in the path o f the file and the filename.


Chapter 2: Getting Started with Predictive Modeling

65

I f you want to pass both the original and transformed variables to the next node, you must change the Hide and Reject properties to No in the Original Variables area in the properties panel and then run the Transform Variables node. I f you open the Variables property in the next node, which is the S A S Code node in this case, you w i l l see both the original and transformed variables passed to it from the Transform Variables node, as shown in Display 2.59. Display 2.59 V a r i a b l e s - EMCODE

Name

Report

Use

Role

Level

credscore

Yes

No

Input

Interval

LOG_credscore

Yes

No

Input

Interval

0PT_VAR1N

Yes

No

Input

Nomina!

resp

Yes

No

Target

Binary

VAR1N

Yes

No

Input

interval

VAR 2 0

Yes

No

Input

Ordinal

VAR3N

Yes

No

l n

VAR4C

No

No

P t Rejected

Nominal

VAR 5 0 VAR6C

res No

u

Type

Order Crt â&#x20AC;˘

rra

Interval

No

input

Nominal

US

No

^Rejected

'Nominal

WE, i

Explore...

OK

Cancel

Help

2 . 6 . 8 SAS C o d e N o d e The S A S Code node is used to incorporate SAS procedures and external SAS code into the process flow o f a project. In addition, you can perform D A T A step programming in this node. The S A S Code node is the starting point for creating custom nodes and extending the functionality o f Enterprise Miner 5.2. The S A S Code node can be used at any position in the sequence o f nodes in the process flow. For example, in Chapters 6 and 7. custom programming is done in the S A S Code node to create special lift charts. The following sections explore the S A S Code node so you can become familiar with the macro variables and macros used internally. The macro variables refer to imported and exported data sets, libraries, etc. First, look at a simple process flow with the S A S Code node, as shown in Display 2.60. Display 2.60 I

En3 Ch2_TranVar

Data Partition

10 Transform Variables

SAS C o d e

Regression

Display 2.61 shows the variables contained in the data sets passed by the Transform Variables node to the S A S Code node.


Predictive Modeling with SAS Enterprise Miner: Practical Solutions for Business Applications

Display 2.61 iVarlables-EMCODE

Use Name Yes credscore LOG credscore yes iOPT VAR IN fes. fes resp VAR1N Yes Yes VAR20 VAR3N Yes VAR4C No Yes MR50 No ^AR6C Mo VAR7C Yes vAR80

I

Reporl

No No No

Ho No

"""NO

No No No No No No

|

Type

Level

Rote

Input Input Input Target Input .Input Rejected Input Rejected Rejected input.

Interval Interval Nominal Binary interval Ordinal interval Nominal Nominal Nominal Nominal Nominal

Order

N N

Label

Credit score Transformed: Credit score TransformedVARIN

c N

N N N C

N

USAGE WEALTH PROW CLUSTER

b

1

'I

Explore... ||

OK |

Cancel

|

Help

To open the SAS Code window, select the SAS Code node in the Diagram Workspace and click on [I]

located to the right o f the SAS Code property in the properties panel, as shown

Display 2.62. Display 2.62 Properly \

N o d e ID

Value EMCODE ...

ilmported Data Exported D a t a Variables ; rToolType

Utility

: rData Needed

No

r-Rerun ••Advisor Type

No

•ilSSwflHflfli

Basic

;-SAS C o d e

:

j - C o d e Location

_ Internal File

L|

;-External File '•[Catalog Entry 1 hScore Code TjScore Code Location

Internal File

1 External File i [-Catalog Entry |-Publish Code ! - Code Format ;

11 Muse Priors l-Time of Creation

Publish Datastep |Yes

j

11/15/06 10:50 PM

; - R u n Id i r-Last Error '•• L a s t Status - N e e d s Updating :

i r N e e d s to R u n h T i m e of L a s t R u n

No Yes

• " R u n Duration 1

"jOrjd H o s t '

L

-I .

The SAS Code window is shown in Display 2.63.


Chapter 2: Getting Started with Predictive Modeling

67

Display 2.63 fttSAS Code ]

Macro Variable

Qirient Value

General

hEM_lMPORT_PATA &E M_LI B. Tr a n s_TRAI N System macro variables. [ â&#x20AC;˘ E M_l M P 0 RT_DATA__C M ETA &EM_UB.Trans_CMeta_TRAIN ! EMJMPO REVALIDATE KEM LI B.. Tra n S~~VAL 1 DATE rEM_IMPORT_VAUDATE CMETAfcEM LIB.Trans CMeta TRAIN i - | {Macro Variables ^r> j Macros pi* J

13

The SAS Code window has three tabs: SAS Code, Macro Variables, and Macros. These tabs can be used to place the macro variables, macros, and SAS code in different locations o f the SAS Code window. Display 2.64 shows the Macro Variables and the Macros tabs placed at the top and bottom o f the SAS Code window.


Predictive Modeling with SAS Enterprise Miner: Practical Solutions for Business Applications

Display 2.64

Current Value

Macro Variable

EMLIB

sasadm kssarma S561 EMWS2

i-EM. USERID H=M, METAHOST i-EM. METAPORT i;EM LIB '•Em PSEP I-EM. CQDEBAR I-EM. VERSION IEM. T00LT7PE I i-EM. NODEID

d Macro variable identifying the workspace library (e g EMWS.)

k

lUTY EMCODE

Macio Variable:

d

SAS Code

EM.DECDATA

I !-EM_REG13TER r-EM_REPORT ;-EM_DATA2CODE j EMJ3ECDATA i;EM_CHECKMACRO i'EM_CHECKSETINIT !EM_ODSLISTON ••EM ODSLISTOFF

Macro used to copy decision data set to WORK and assign the proper data set type (PROFITA-OSSyREVENUE). The arguments are NODElD-node identifier, D E C D A T A - decision data set, D E C M E T A - decision meta data set

d

Macros ; ; t:

In Display 2.64, i f you highlight the macro variable E M J d B , its description appears in the right panel o f the window. By scrolling down inside the Macro Variables tab, you can see the macro variables identifying the data sets imported from the predecessor nodes, as shown in Display 2.65. Display 2.65

Macro VeiiaUa

Currari Value

B : • E M_IMPO RT_DATA I-EM IMPORT DATA CM ETA I-EMJMPO RT_VAU DATE I-EM IMPORT VALIDATE CM ETA l-'EM IMPORT TEST r E M_l MP 0 RT_TE ST_C M ETA i-E'.'_IMPORT_SCORE I EM IMPORT SCORE CMETA

6EM SEM SEM SEM SEM SEM

E M _ I MPO RT_D ATA

d UB Trans UB,Trans LIB Trans UB Trans UB Trans LIB Trans

Macro variable

TRAIN CMela TRAIN VALIDATE CMeia TRAIN TEST CMela TRAIN

identifying the training data set.

d

Macro V i i i i b l e s

In Display 2.65, the macro variable E M _ I M P O R T J D A T A is highlighted. Its current value is & E M _ L I B . T r a n s _ T R A I N . The first part o f the current value is the macro variable & E M _ L I B whose value is E M W S 2 (see Display 2.64), which is the libref The second part is T r a n s , which indicates that the data set is created by the Transform Variables node. The third part is T R A I N , which means this data set is the training data set. From the description given in the right panel o f


Chapter 2: Getting Started with Predictive Modeling

69

Display 2.65, it is clear that the highlighted macro variable refers to the training data set. By scrolling down farther, you can see the macro names identifying exported data sets, files, number o f variables, and code statements. Display 2.66 shows some sample SAS code. In this code, I created a composite variable and included it in the exported data set. Display 2.66 - ilata

6en_expor.t_tr.ain; s e t sem_impor.t_data ; t o t a l _ t r . a n = var3n + v a r l n ; l a b e l t o t a l _ t t a n = Tot*2 Transactions

ran

Created

in SAS Code node"

;

;

a SAS Code

111 the sample code shown in Display 2.67. I create a new variable called total_tran. which is the sum o f Var3N and Var I N . Var I N and Var3N represent the number o f withdrawals and deposits from a bank per month, respectively. The composite variable total_tran represents total transactions, and it captures the customer's overall activity. The next node after the S A S Code node in the process flow is the Regression node. I f you select the Regression node, and click on [7].

located to the right o f the Variables property, all the

variables including the newly created variable total_tran appear as shown in Display 2.67. Before you do this, you must update the path. Display

2.67

... Variables-Reg2 Name

port

LOG credscore OPT VARIN VARUM V*R20 VAR3N VAR4C VAR 50 VAR6C VAR7C VAR 80 credscore fesp total (ran

X] Role Input It) put Input Input Input Rejected Input Rejectad__ Rejected input Input Target input

Level nterval Nominal nterval Ordinal nterval Nominal Nominal Nominal Nominal Nominal Interval Binary Interval

Type N

0

N C

C N N W N

Order

Label Transformed: Credit score Transformed VARi N

USAGE WEALTH PROXY CLUSTER Credit score Total Transactions Created in SAS Code node

_il Explore..

L

OK

Cancel

Help

The S A S Code node can be used with code located in an external file. To use an external tile, select the S A S Code node and set the Code Location property to External File, and set the External File property to the full path o f the IIIe and the filename. This is shown in Display 2.68.


Predictive Modeling with SAS Enterprise Miner: Practical Solutions for Business Applications

Display 2.68

r

Value

Property EMCODE2

N o d e ID Imported Dala Exported Data

.. J

Variables

• •

•iToolType

Utility

• Data N e e d e d

[No

•Rerun

iNo Basic

-Advisor Tvoe

__

-;SAS C o d e -Code Location

External File

- External File

C:\TheBook\EM

5.2lProgramsU_TV.SAS

- C a t a l o g Entry -Score Code Score C o d e L o c a t i o n

Internal File

-External File - C a t a l o g Enlry -Publish Code

Publish

-Code Format

Datastep

• U s e Priors

Yes

• T i m e of C r e a t i o n

11/16/06 3:01 PM

k

R u n id .Last Error - L a s t Status -Needs Updating •Needslo Run

Yes

- T i m e of L a s t R u n -Run Duration -Grid H o s t

2.6.9

Drop Node

The Drop node can be used for dropping variables from the data set or metadata. After an initial exploration o f your data, you can drop any irrelevant variable in this node by setting the Drop from Tables property to Yes. This setting w i l l result in dropping the selected variables from the table that is to be exported to the next node. Display 2.69 shows the settings o f the Drop from Tables property.


Chapter 2: Getting Started with Predictive Modeling

71

Display 2.69 Property

Value Drop

N o d e ID I m p o r t e d Data Exported Data Variables j- Drop f r o m T a b l e s

Yes

[•Assess

INo

[- C l a s s i f i c a t i o n

INo

K ^

! [-Frequency

INo

[hpdden

Yes

r- Input

:No

[•Predict

iNo

1* R e j e c t e d

Yes

[•Residual

No

[-Tarqet

[No

-Other

No

[•Time of C r e a t i o n

1 1 / 1 5 / 0 6 10:50 PM

;

[ R u n Id [•Last Error [-Last Status [•Needs Updating

Yes

[ • N e e d s to R u n

Yes

[ • T i m e of L a s t R u n [•Run Duration i =-!Grid H o s t Drop f r o m Tables I n d i c a t e s i f t h e s e l e c t e d variables s h o u l d b e d r o p p e d f r o m t h e c r e a t e d tables i n a d d i t i o n t o d r o p p i n g t h e m f r o m t h e metadata. I f set t o Y e s , t h e n t h e n o d e w i l l c r e a t e data sets i n s t e a d o f v i e w s . I f this p r o p e r t y is set t o N o , t h e n t h e v a r i a b l e s a r e o n l y d r o p p e d f r o m t h e e x p o r t e d metadata.

To select the variables to be dropped, click on ___ , located to the right o f the Variables property o f the Drop node. I f you want to drop any variable, click in the Drop column next to that variable and select Yes, as shown in Display 2.70. Display 2.70 __J Mime credscore LOO_credscore OPT_VARiN resp VAR1N VARIO VAR3N VAR50 VAR6C •AR7C VAR BO

Drop Default Default Default Dirfautl Default Del-* DetauB NO lefault Default Default <l

Role Input iroui Input Target input Input Input Rejected Input Rejected Rejected Input

L-.-veH Interval Interval Nominal Binary Interval Ordinal Interval Nominal Nominal Nominal Nominal Nominal

. Type N N C N

Order

Label Credit score Transformed: Transformed

N

IN USAGE WEALTH PRC CLUSTER

N

jc N

1 Expk*e

OK

Format

m


72

Predictive Modeling with SAS Enterprise Miner: Practical Solutions for Business Applications

2.6.10 Decision Tree Node The Decision Tree node is discussed extensively in Chapters 4 and 7 using several examples. In these chapters, the Decision Tree node is used to develop models with binary and continuous targets.

2.6.11 Neural Network Node This node is covered in Chapters 5 and 7, where I develop models for predicting response and risk for an automobile insurance company, plus price sensitivity and attrition for a hypothetical bank.

2.6.12 Regression Node This node is discussed in Chapters 6 and 7. Models are developed for binary and continuous targets.

2.6.13 Model Comparison Node You can use the Model Comparison node to evaluate a model or to compare models using your test data set. Alternative models to predict risk and response are evaluated in Chapter 7 using the Model Comparison node. Chapter 7 includes a detailed discussion of lift charts and capture rates in the training, validation, and test data sets for each model compared. I f you want to evaluate a model or compare models using the test data, you should use the Model Comparison node.

2.7 Summary •

This chapter introduced you to Enterprise Miner 5.2.

After reading this chapter, you should be able to open Enterprise Miner 5.2, define a new project, and create data sources for the project.

The chapter gave an overview of the tools (nodes) that are available for structuring and customizing data mining tasks. It provided information to help decide which tools to use and which values to assign to the various properties of the tools in order to perform data cleaning and exploration tasks for the data mining project.

Some of the tools mentioned here are covered in depth in subsequent chapters.

There are several other tools in Enterprise Miner 5.2 such as Cluster node, Principal Components node, and User Interface node, which are not covered in this book. Y o u can find information on these in the S A S Enterprise Miner help and documentation.

2.8 Appendix to Chapter 2 2.8.1 Cramer's V Cramer's V measures the strength of the relationship between two categorical variables. Suppose you have a categorical target Y with K distinct c a t e g o r i e s ^ , Y , and a categorical input K


Chapter 2: Getting Started with Predictive Modeling

X with M distinct categories X , X , ....X }

X,

2

M

73

. If there are n observations in the /'* category in v

h

and j' category in 7 , then you can make a contingency table of the following type:

X\Y

Y2

YK

Row Total

".2 n

n

u

\\

n

2K

n

i.

•• n

M2

Column Total

n

N

MK

N

2

To calculate Cramer's V , you must first calculate the Chi-Square statistics under the null hypothesis that there is no association between X and Y. M

K

X- = Yl}""

f'

j )

, where E =n ¥

X

Cramer's V =

J

%

, i=\,...M

and j = \,..K.

, which takes a value between 0 and 1 for all tables larger

/Vmin(A/-l,/:-l) than 2x2. For a 2x2 table, Cramer's V takes a value between -1 and +1. Suppose that T i s a binary variable with two levels, and X is a 16-Ievel categorical variable. Then Cramer's V = J ^ - .

VN

2.8.2 Calculation of Chi-Square Statistic a n d Cramer's V f o r a Continuous Input As pointed out in Section 2.6.2.1, the A G E variable is divided into five groups (bins). Table 2.1 shows the number of responders ( R E S P = 1) and non-responders ( R E S P = 0).

Table 2.1 RESP Age Group 18-32.4 32.4-46.8 46.8-61.2 61.2-75.6 75.6 - 90 Overall resp rate

0

1

3830 4952 6915 3882 946 20525 0.686363

1959 2251 3181 1597 391 9379 0.313637

Total 5789 7203 10096 5479 1337 29904

Table 2.2 shows the first step in calculating the Chi-Square value. It shows the expected number of observations in each cell under the null hypothesis of independence between rows and columns.


Predictive Modeling with SAS Enterprise Miner: Practical Solutions for Business Applications

Table 2.2 RESP Age Group 18-32.4 32.4-46.8 46.8-61.2 61.2-75.6 75.6 - 90

0

1

Total

3973 4944 6930 3761 918

1816 2259 3166 1718 419

5789 7203 10096 5479 1337

The expected number of observations is obtained by applying the overall response rate to each row total given in Table 2.1. 2

(O â&#x20AC;&#x201D; ES , where O h the observed frequency as

The Chi-Square for each cell is calculated as

E shown in Table 2.1, and E is the expected frequency (under the null hypothesis) as given in Table 2.1. The Chi-Square values for each cell are shown in Table 2.3.

Table 2.3 RESP Age Group 18-32.4 32.4 - 46.8 46.8-61.2 61.2-75.6 75.6 - 90

0.0000

1.0000

5.1722 0.0134 0.0304 3.9202 0.8748

11.3187 0.0292 0.0666 8.5789 1.9143

These cell-specific Chi-Square values can be retrieved from the Results window of the StatExplore node by clicking View->Summary Statistics-^Cell in the menu bar. These cellspecific Chi-Square values are shown in Display 2.71.

Display 2.71 Target resp resp [resp jresp resp resp resp resp resp resp

| AGE AGE |AGE AGE AGE AGE AGE AGE AGE AGE

Input

\ Target For... (Input Forma.I Frequency... {target Num...| trtptft Nume...|j CM Square f Log Chi Squ 3830:

18:32.4 32.4:46.8 46.8:61.2 61.275.6 75.6:90 18.32.4 32.4:46.8 46.8:612 61.275.6 75.6:80

7

0 J i 49S2 Z,.f o fZZZM 691& ' 6 47: 3882* *_* f\_ ' 0 62: 946* 76; 0 1959 _ 1 '18 2251 " " i ; " 3 3 : 3181 " "47 " Z.Z7A 1597 _ ~ 6 2 Ji " 391 '* r 76: 8

!

:

2

By summing the cell Chi-Squares, you get the overall Chi-Square {x )

5.172157} 0.030431 3;920158 0.874759 11.31875; 0.029237! 0.066592 8!578872! 1514323]

1.64329; ^431549! ~-3.49234j 1^366132} -0.13381 r 2.42646: -353232 i

-2.709171 2149302j 0549364

of31.9864 for the A G E

variable. Since there are 29904 observations in the sample, Cramer's V can be calculated

| ~i 29904

which is equal to 0.03267.

By selecting the Chi-Square Plot (or clicking in it), and selecting the Table in the Results window, you can see the Chi-Square and Cramer's V for all variables. A partial view of this table is shown in Display 2.72.


Chapter 2: Getting Started with Predictive Modeling 75

Display 2.72 R e s u l t s - 5t a t E x p l o r e

:

File

Edit

View

•Jo]

a

1

Window

]_!*!

Chi-Square Plot - EMWS,5tol_CHTMEJ

SEGMENTID

T=.rget

[OVERALL.

resp

_OVERALL_ _OVERALL_ _OVERALL_ OVERALL_ .OVERALL. _OVERALL_ _OVERALL_ _OVERALL_ _OVERALL_ ^OVERALL, _OVERALL_ LOVERALL, _OVERALL_ _OVERALL_ OVERALL

resp

L

F

<l

resp

resp resp resp resp resp resp resp resp resp resp resp resp resp

Input RESTYPE MFDU NUMTR MRTGI

Cramer's V

Chi-Square

Prob

0.165349

817.5806

= 0001

0.162414

788.8147

<00Q1

0.13795

569.0762

<.0001

0.112498

378.4571

<.0001

INCOME

0.06085

110.7257

<0001

AGE DELINQ CRED

0.032671

31.9186

<.0001

0.03132

29.3346

0.0001

0.027085

21.9140

<.0001

EMP_STA

0.022864

15.6326

0.0004

MILEAGE

0.020033

12.0011

0.0173

HEQ

0.019453

11.3215

0.0789

RES_STA DEPC GENDER MS

0.016075

7.7274

0.0054

0.011413

3.8952

0.0484

0.008698

2.2624

0.1325

0.007163

1.5366

0.4638

MOB

0.002769

0.2294

0 6320

I


^ 3 ^ 1 3 ^ 1 0

3 0 1 0 0 1 1 0 1 1

-

,

3.1

|||1|^|||1011011

3.2

^31*131116 8

3.3

7|^1118|01|11181!011

3.4

3 0 1 1 1 1 1 1 8 1 7 誦 隱 ^ 霸 ' ^ ― ― ― ^ … " 画 漏 漏 , … ^ 漏 園 … 屬 … " 棚 纏 " ^

6

1 0 #

"

3 1 1 1 1

^ ^

霸 0

^ 1

^ 1

纖 0

^ 1

聽 1

… ^

" 龜

漏 ^

7 7 8

^ 8 1 * 1 & ― . . . . . . ^ 霸 ^ 漏 " … … " 編 ^ 漏 " … 9 9 1 0 9 1 1 0

3.5

3 . 1

7

1111^01111011011 丁 1115 0卜3|316『^111 ^1X111131*126 乂00

1卜6 5八5 601617)1*156 1^4106!"

&1)1& 86160(1011 &1\6

1^308『0『111 ^31*131)168 0 0 ^ 6 5 5 \^111011 31*6 655601131『0『51111(1111^ ^6^1011^ 86160(1011 00(16 15 0561\11

丫011

1^06^\5^ 丁116 ^31*13516

10 1(16011^ 10115011&01 ^31*1&6165 (^!?口氹)?0「 1)011(^05

131-6(11011^6 ! ^ ^ ^ ^ . 丁 116 XI08101-111 ^31131)168 110016 151*0^(165 &

&

&^/ 0^ 1X305^0111131:10115

1\\31 03.11 60 3^13116(1 10 1116 111^11(5『0『111113^0^1118化6 ^1^60131011 0 ^ 1:116 口『6出01卜6 0103615. 丁 116 ^ 0 1 1 0 ^ 1 0 2 10^)105 21!"6 00^61*6(1: 肇 ^ 3 0 3 1 ) 1 6 86160(1011 00(16 0

^31*13616 3616011011『0!" 03(0301*1031 ( ( ^ & 1 3 ^ 6 1 5

0

^『131)16 561601100『0『000110110115 ( ^ & ! ^ ^ & ^ ) 131^615

0

1)111111112 000110110118 111)311(3『0『^31^1)16 361601100

0

3『0叩1113 0「 00113口510@ 1116 031:620^165 0 ^ 03(0301*1031 111|3111;5『。『^31*1&1)16 501601100

0

8 八 8 00(^6 2)^X1^^3.1&6 !)乂 1)16 ^31-131)16 86160(1011 00(^6


78

?^101\^6 ^[。(^""宕^1(17 8/48 ^11(6^186滅"":?^00^100180114^10^15^0!"

9

8148111658 ^?"^!。"3

X 1*3118101*111 ^31*131)168 110(^6 0

111^116111^1031 11^115『011113110115 0『0011(111110113 !111311(3

0

13111111113 0011(111110115 ^31*131)165

0

01*6311113 ( ^ 爪 爪 乂

0

1 1 0 ^ 10 11131(6 0101*6 1)1311 0 0 6

0

5八5 006& ^61161-3(601 5^ 1卜6 11*308(01*111 ^31*131)168 110(16

&6165『0『0316301*1(^1 ( ( ^ & ^ ) 111(31115 0『^05^01111311011 0 『 3 0 1111)111

,1151113 1116 1 ^ 0 110^65 !11 3116171311^6 56^11611008 0

10 301116 03565, ^31110111&1*1^ \^1160 11161*6 31*0 & 1 3 ^ 6 0110166『0『1013015, 111113^ 1)6 1565110 111&1(6 1116 乂3『!3516 56)60(1011 111*81: 3.1)6 111611 ^ ^ ? ! ^ ! ! ! ! 11)6 56160(6^ 乂3『13|3165,

0

III 01116『03563 & 0 1 3 ^ 1)6 1)65110 1136 1116 丁!^!^&!^!!! ^31*1&1)168 110(16 ^91*13516 861601100 1X0680111111656

III 11113 56(^011, I ^1301133 110\^ V & ! ' 3 1 * 0 56100(6(1 ^ 111638111*61110111: 30&165 0『1卜6 111)311(5 \11 口『601101~6 1110(1611113

111611 1(16

36^11611063 31*6 11111511*316(1

1116 ^31*1&1)16 86160(1011 110(16『0『^1^61*6111:

1116 131^61, 311(1 I出501155 1卜6 ^ 3 1 ^ 1 ) 1 6 3616011011 0『116『!1

1111111113

31*6 0 ^ 6 0 0001x0016(1

!^!!!!^!1 0 ^ 1111311(5

&

\ ^ 1 \ 0 ^ 0 ^ 雷 丁 116 口"!!!!^ 0『130(6111131 1叩"15 10 0^10056『「001 0 1 3 ^ 1)6 3 8 131*^6 3 5 2 0 0 0 0『111@|16^ 30016 0『111636 1(11)11(3 0 1 3 ^ 001113乂6 3 0 ^ 1*6131101110 1卜6 1 3 ^ 6 ^ 八 I I 101(131 301*60111113 !5 1116『6『0『6 0 6 0 6 8 5 3 1 7 化 6 1 1 1 ^ 1 ^ 6 111*616^4111;

&1)165 10 1(66^ 1116 11111111?6『0^ 1111311(3 10 & 1113113^1)16

5126.八『11131 3616011011 0311 111611 1)6 1113^6 & 0 0 1 出 6 1*6111&1111113 ^31*131)163 1131113 6

&

3 516^156

1*631*6531011 0 『 & ^60131011 1『66^ 丁卜6 ^31*131)16 86160(1011 110(16 1361^011113 1^16 1111(131 861601100 38 ^ 6 1 1 3 5 3 ^ 3 0 3 1 ) 1 6 36160(1011^ I XVIII 011301153 1116

1)16 ^31*131)16 8 6 1 6 0 ^ 1 0 0 00(16

出65

0011111131~5

0^1031~

50316(1 !11口1115, 8 8 \^611 & 8 1116 36160(1011 0『1祕『!&『0『0310801*1031 311(1 0011(111110115 13^613^ 丁卜6 0113011351011 \^\\\ ^00115 00『00『001X11X100 51(113(10115 ^ ^ ^ ? 1-4^: 鲁 丁 116 131-^61 15 0011(111110113 (!!!^^!-?^&^(!) 311(1 1116 1111)11(8 31*6 011016!"1(1, 0011315(1113 0卩 1 ) 1 ( 6 1 ^ 3 1 - 3 0 3 1 6 ( 1 ^ 0 & 5 0 1》 #

9

丁卜6 131*^61 13 0011(111110113 311(1 1^16 1叩1115 31*6 03(6^01*1081 311(1 !^!!!^""^. 711636 31*6

7116 1 3 ^ 6 1 15 1510&1^

肇 丁 116 131*^61 !5 1)111317

#

^ 6 1叩1115 31*6 0110161 10 10161^31-3031601 ^ 3 5 6 3 》 1116 1111311(3 &1*6 0316201*1031 311^ 00011031-3031601 ^ 0 & 5 6 ^ ) .


0^006^

3: 1^01-101)16 86160(1011 017(1 7^(1^011^011011

0^ 「^/^/&5

79

丁。『00115 0 0 扒 0 ∼ 6 3 0 1 1 0 3 8 6 18 113)1(1163, I 1 1 ^ 6 1156^ 56(331^6 6&X& 36(3『0『111113(1^(1011 |3111^0563^ III 3(101111011, 10 ^66(5 1116 0(13011381011 @61161^1? I 113^6 111(1103(6(1 &11 111111161110 11^1115 35 !^乂八!^, 1^乂八!12, &0(1 311 000110&1 ( ( ^ ! " ! 0 声 5 33 0 ^ 八 1 1 1 , 0\^八112, 610. ! " 3 1 * 6 501116 6X3111^163 0 『 3 00111111110115 131*^61: ,01131136 10 53^1025 0(6^05115 0 ^ 3

'!50111611031 1)31^

& 1*316 0(131136

參 011510016『,5 35561 乂31116 0^111001 1x001 3 5111^6^ 0『& 1)1*0^61*^6 00011)30^ 參 !053 116^116110^ 0 『 3 0 311(0 1115111*31106 0001卩30乂

八II 0『1(16 6&X&

1 1136 1161*6 15 &II 51瞧1^6(1, 3 0 II ^ 0 6 5 0 0 1 V V 10 3 1 ! ) ' 1 ) 3 1 ^ , 610.

丁 116 1)10317 131*361 1156(1 III 1)16『0110\^113 6X31111?|63 ^111 「6(31*656111 3『65口01156 10 3 6\1601 03111^1311 ^

& 11^|301|1611031 311(0 1113111^1106 0001口30乂. 丁卜6 610&1^ 1 ^ 8 6 ^ 3

5

& 5 03016 511886515,

131(65 001乂 1\^0 乂3|1165: 1*65)301136 (^^^^^^^^ 6 ^ 1人30(1 0 0 1*63(301136 ( ^ ! ) ^ ^ ^ ! ! ^ ^ I?乂 0丫 III ^!!^!^^!?^ \1;116『,1)0(11 0^111&I 31101 11011111131 1111311(3 31*6 ^3.1&6

3.2.1

3 8 01385 V^I*1^^^163.

0 0 1 1 1 1 1 1 1 1 0 1 1 3 7 9 1 ^ 6 1 : 1/^1111 111111161*10 1 1 1 ( 6 1 1 ^ 3 1 - 8 0 3 1 6 ^ 画 叩 1 1 1 5

III 11113 6X811113169 & 11乂口0111611031 1)311^ \^31115 10 111635111*6出6 6 ^ 6 0 1 0 『 & 151*0111011011

!1^0^65 311

11101*6356 III出6 111(61*631: 1*3(6 0^61*6(1 0 0 53乂||135 ^6(3081(3 III出6 \3311^ 丁 116『6『0『6, 1卜6

乂3『!3616

13 1116 0(181136 III 3 0115100161- ^ 016^)051(5 !11 1*65(301156 10 1(16 !110116356^ ∼!1611 1)16 131*861: 15 0011(111110113 ( ( ! ! ^ ! ^ & ! - ^ ^ ! ^ ) , 1卜6 ^31*131)16 86160^1011 110(^6 11363 11-3(1 II&1*63 36&1111 0「&61*1011 0『56160(1011 ^

63011 10161^31-50316(1 11^1111116 \^3^\^

0^10^13(63 1 ^ 0 111633111*65 0? 000*6131100 1)61^660 63011 111|3111; 311(1出6 1)61X^660 1)16 131^61 ^ 出 6 01*1^11131 ― 化 . 7 1 1 6 1)111116(1 乂6『5100 0『1)16

5

^6

86160(1011 110(10 0)16 15 1)16 11-8^1131*6

13 ([!6 11-8^1131-6 1)61^600 1116 131*^61 30(1 1116

^&. 丁 116 1)111116(1 乂3『181316 13 & 08(^01*1031 乂3『13616 01*631601 \^ 1116

^21131)16 861601100 110^6 &0111 63011 0011(111110113 ( ; ! ^ ! " ^ ^ " " ) 1叩111 丁卜6 16乂6|5 0『11115 03(6301*1031 ^21*13616 31*6 1116 5105. III 50161731*156 \ 1 | 0 6 ^ 1)113 6111116(1 ^31*131)16 15『6化1X6(1 10 38 311 八 0 \ ^ 1 6 \^\2^&.

丁 116 ^!!!!!^!' 0 ^ 16^613 0「 031630『;65 0『1卜6 1)1011601 ^31*13616 ^ 0 ^ \ ^

16, 001X631)011(1108 10 1 6 10161^318 111&I &1*6

13 3.1 01031

!0 ^1(1111# 11115 ^ 0 6 3 0011116&II 111311116 ! ^ 卬 ! ^ ! 1 0 『

1*6001X15 III 63011 1)111 15 &0(11^1 丁0

& 1)6116!" 1310(111*6 0『1)1636 1)111116(1 ^^131?1639 161 05

3 0 6 \ & & . 311|3|3056 ^ 6 31*6 1)111111113

1 1 1 6 八 0 6 乂3『13616, &0(1 II 1^1^63 0 0 乂31065『『001 1 8 - 8 1 ^ 丁 116 111*51 1)111 100111(165 &11 !0(11^(111&15 "!出 3^63 18̶21,出6 36001101 " 油 2 2 - 2 5 , 311(1 5 0 011. I卩1(161*6 3『6 0 0 03565 111 3 ^ ^

6 0

出 3 1 1)111 13

311(011121110311^ 6110110&16(1, 3 1 1 ^ 出 6 011111|36『0^61115 ^!11 1)6『6\^6「 11130 1 6 . 7116 6111116(1 ^31*13516 & 0 0 1 八 0 6 13 0&]]&6八0乂16一八06,

7116 !?!"!!" ^\^\6

^

11^100\15 15 08116(1 ^ 0 ^ 1 6 ^ 0 0 1 ^ 1 ^ 30(1

3 0 011. 1(1 1110

& 561 1156(1 III 11115 0|13^6「,出6「6 31*6 2 3 0 10161^31-80&1601 1111311(5 311(1 1 1)10317 !0口

80016

0『1(1656 10^1^31 ^31*131)165 \^6「6 0「1^0311 乂 0『出031, 1)111 I 000^6116(1 111601 111(0 10161^31 30&16. 3 ^ 6 3601100 3.5,"八口口60出乂 10 013|516『3,,,『0「 ^31*135165 III & 63\3^ 5 0 1 1 1 ^

0 0 ^ ! ! ^ ! ! ! ^ 出 6 016350『6111601

0卩


?1-6(^1011^6 ^00(61111^

""/? 848 ^17(6^186

1^11761": ?^00(1001 80(11(10118/。丫 8118117688 ^^1100(10173

7116 2 3 0 ^3021)165 &1*6 03016(1 X 乂 八 尺 0 0 1 , 1^乂八!1002, 610. 7116 ^31*231)16 861601100 1106& 01-6316^ 2 3 0 1)111116(1 461

&1)163 ( 八 0 乂 1 6 V3I^13^)1^8) & 0 0 1 1116 101:61^31-50&16^ V&!"1&6163, 30(1 1)1611 1110111(16(1 &11

&6165 111 1(16 3616011011 口『00651八I ^ 6 611^ 0『出6 \^^13516 5616011011 口『00655, I 113(11^0

01*1811131 ^ 『 1 ^ 1 6 5 , ^ ^ 1 1 1 3 3 八 0 乂 16一∼乂八11049,

1^1乂八11239, ^61卜『66 61006(1 ^1^1?165,八0乂16》乂八11133,

八0乂 16一XV八11239.

~

丁 116 10110^1118 15 3 016301^1131:1011 0 ^ 出 6 ^31*136168 ^3.1 ^61*6 36)60(6(1: !^乂八11133狄618

6

1(1 1116 00510016「,5 0)1001(1113 &000001『0『1卜6 12-0100111 ^61*10(1

^10110

1116 |31^01110110031 11101*6336 111 1116 11116『6311*310 &11(16 11^(30111611031 1^乂八11049 00011)60『II11330(10113 1^\^八11239

1116 I&51 1 2 01001113 6乂 1116 0051:01116『

016^0311; 5&1^1106 3 01001113 |3『50『10 1116 ^011101;101131『&16 0\\31\^^

-

1^016 11131 ^31*1315165卜乂八III33 &0(1 X V 八 1 1 2 3 9 31*6 3616016(1 111 1)0111 1^6^ 01*1311131 &1111 &0(1 III 111611

5 1 1 1 1 1 6 ( 1 八 5 ^5011556(1 111 0113^6!" 2, 001 乂 八 0 乂 1 6 ^^1315163 31& 133536(1 10 1)16 116X1 110^6 1^乂00 5611116 乂3106 0 ( 1 1 8 6 人 0 ^ 1 6 ^31*1&1)168 ^0^6^ 151-0^611^ 10

10

八 5 I 1 1 ^ 6 3611116 1 1 8 6 入 0 \ ^ 1 6 ^31*135168

111 1)115 6X311!]?!^, 1^乂八11133 &0(11^乂八11239 31*6 0 0 1 ^3836(1 10 1116 0 6 X 1 1106& 35

;叩"13. 出 ^ 化 站 ^ ! 111 011&口16『2, 56^1011 2.6.6.1, 1116 11~80|113^6 1116(110(1 11563 & ^ 0 - 3 1 6 ( 3 3616011011

八5

^00&6\1^6^

1)1 1116 111*51 516口, & 131*61111110317 3616011011 13 1X1&6&, 1)356(1 0 0 1\110111111111 ^ - ^ ^ ! ! ^ ^ . ?0『

1116 01*1311131 V^'!&1)165 1116 11-8(1113!"6 13 0310111^(6(1『1*001 & ^ ^ 6 3 3 1 0 1 1 0 ? 1116

0 0 63011 !0|301;『0『

1116 1)111116(1 ^31*131)105 1! !5 0^10111&16(1 & 0 0 1 & 0 0 6 - ^ 3 ^ 八 0 3 1 ^ 5 ! 5 0^\^31^13006 ( 八 1 ^ 1 0 V 八 ) . 7116 ^6&1111 ^ ^ 1 0 6 0 ^ 1116 1^111111111111 11-8(11131*6 ^ 0 ^ ^ 13 0.005, 1)1111116 11361- 0811 5611116 11111 ^8(11131*6 口『0口611^ 10 & 出 卩 & ! ^ ! ^ 乂311165 &1)0^6

1 『 ^ ! ^ . 丁卜6 111^)1118 11131 1 1 ^ 6 11-301II&1^6

\\\6 10∼6『1)01111(1 561 6乂 1)16 1\110111111111 8-8^1131*6 ^0^6^

3『6 !110111(16(1 111 1116 30000(1

8(6)3 0『1116 V&!'131)10 56160(1011 口「0。655, III1116 360011(1 516口, & 30^11011(131『01^31x1 56160(1011 |3『00655 15

丁 11!5 [?1-00655 513115 I?乂 36)6011113

1116 111^111: ^31131)16出^ 1138 1116 111^11631 000*6131100 006^101601:

1)16

6^113(1011

( 爪 ^ ^ ^ )

15 63(1111&16(1∼&卜1116 36160(6(1

八1^6^6381011

八 I 63011 511006331^6 3(6)3 0『1卜6 56(^1161106, 3 0

3(1(111101131 111(3111: ^31*131)16出31 !31*0^1(165 1116 13『36对11101*6111611^1 000^1131111011 10 1116 ^10(161 ^ 8 ^ 1 1 3 ^ 6 13 &66&6

1 0 出 6 1-681*6531011 ^ 1『伤6 1。\^6『1)01111(1『0『^6 11101*6111611(31 000(1*11311(1011 10 1)16 ^10(161

11-8(111&!"6 13『630116(1, 1116 36160(1011 口『00655 510口5, 7116 10∼6『1)01111(1《0『1116 1001*601611131 00011*11)1111011 10 1116 1^00161 ^ 8 ^ 1 1 3 ^ 0811 1)6 3(3601^16(1 ^

3611108 ^16 810口 ^8(11131*6 ^0^6^

\0

1116 ^6511*6(1 ^ 1 1 1 6 , 10 11115 6X3111口 16, 1116 1^10111111111 ^~8^11!1^6 ^1^0|3611^ 15 56110 0.05 311^ 11)6 8(011 ^ - ^ 1 1 3 1 ' 6 ?^013611^ 15 561: 10 0 . 0 0 5 , 10 11115 6X3011316, 9 2 乂虹!&1)165, 50016 01*1^11131 311^ 50016 ^!!!^^!, ^61*6 36160(6(1 111 ^11-51 316|3 30(5, 0『111656, 0 0 ! ) ' 6 乂 6 V & ! " ! 3 1 ) 1 6 5 3 6 1 6 0 ( 6 ( 1 ! 0 1116 360011(1 516口^ 丁 116 311131161~ 1(16 ^^11165 10 ^ 1 0 1 1 ^!0111111111 1^-8(111&!'6 311^ 8(011 ^8(11131*6 |3『013611;65 31*6 561, 1卜6 1 3 『 8 『 ^ 6 !^!!!!^!" 0 ^ 6

^31-1&1)163 111(0

&1?165 ^3856(110 1116 116X1110(16 ^1111 1)16 & 0 1 6 0^10|111^ 1.6111118 0101*6

0 6 X 1 110^6 013^65 010『6 361136 1^ 11161*6 13 & 1)6018100 71*66 0『^631*688100 0 0 ^ 6

(!。∼!^!!^^!!!, 35 1)0111 ^ 6 3 6 110^63 0&II 1116111561^68 |361^0170 1*1201*0113 0161110(15 0 ^ 5616011011. ^31*131)16 501001100 15 &11(01113乜0 111 1116 1)60181011 丁1*66 110(16 1)111; 066(15 10 1)6 6X13110111^ 3^601110(1 111 1^16 0356 0『 1116 ^6^1*6881011 0 0 ^ 6 ^


01\00^\~ 3: 1^0!'10^16 86160(10/7 011(1 丁"0113/0"賺"017 0^ V。"!。1)165

丁0 1156 1116

5616011011 00(16『0「

1116 8 1 0 ^ 11-8^1101*6 |3『0|361^ 10 3

10110317 ^31*13616 561601100 00!乂, 561 1116 111111111111111『0『

1 0 ^ 乂3|06 511011 35 0 . 0 0 0 0 1 . XII159 111 6伃601, 6乂口35565 316)3 2

0『1116 ^31-131)16 5616011011 13^00655? ^111011 15 0 ^ 3 ^ 511106 316(3 2 0 3 0 3 1 ^ 3 ^ 5 1)6 1361^011116(1 111 1116 1)60181011 11*66 00(^6 0『1116 ^6^1*6881011 0 0 ^ ^ 0

丁 116 11-3(111&1*6『 『0「1^031『3∼10|31115 0 3 ^ 1 1 1 ^ 5 出 6 1!063「 「61&11005111 口 1 ) 6 ^ 6 6 0 出 6

31101

1116 0 0 5 1 0 3 1 10)3111, ^ 1 1 6 1卜6 11-3^1131-6 1 ) 6 ^ 6 6 0 3 6 1 0 0 6 3 ( 八 0 乂 1 6 〉 ^31-131)10 31\6 1116 !31^61 03^111*65 1116 000-1;063『1*61^101151^口. 1|1 00『6乂3111|3|6, \^八11049 ^ 3 5 0 0 1 56160(601 111 1(5 01*1^11131『01111, 1)111 1116 1)10116(1 乂61^!00 0『II, 0311161^八0乂 16一1^乂八11049, ^ 3 8 5 6 1 6 0 1 6 丄 8 ^ 3

『0『63011 5111 ^ '

1 1

^出

6

^ ^ ( ( ! ) ! ^

1)111116(1 乂3「131316, 乂011 0311 366 1卩1卜6「6 6X15(3 3

1116 1 3 ^ 6 1 1116311 1*613110115111|3

1)61^660 1116 10(3111 &1\6 1116 1(1 1116 ^ 0 1 1 0 ^ 1 0 2 ^ 3 @ 6 5 ? I 5110\^ 1 1 0 ^ 10 561 11(3 ^ 6 口『0口61^65 0『1(16 ^21131)16 86160(1011 110(16 111 5 0 1 6 ^ 1 5 6 1^1106「 5.2, 0!5口13^ 3.1 3110^8 1116 ^1^00655 110^『0「 1)16 ^31*131)16 3616011011^ 018(3187 3.1

"!^81^41战0 03(3 ?&1\\1\0^ 36160(1011

^69^633100

丁 116 6『511001 111 1116 ^1^00653 1 1 0 ^ ( ^ ^ ^ ^ ) ' 3 - 0 15

00(^6, \^111011 1*63(15出6 5 八 3

1110(1611113 63.1& 361. 丁 116 ^3(3 501 1 5 出 6 0 138111110116(1 11110 1^0111111^ ^?//^///。〃, 30(1 1681 ^3(0 5615. 丁 116 8616011011 1100(6 36160(3 乂3「1&1)165 1151(13 001 乂 1116 ^31111113 6.2X2^ 561. 丁116 5616046(1 ^31-131)165 31*6 1^1611 口355601 10 3 1(6^1*6881011 110(16, ^11)1出6 ^ 0 1 6 0?111口111 丁 116 36160(1011 0『!16『100 15 5(360111601 I?乂 56(11113 |31^0|5611165 口3061 0『1116 \2^\2^\^

^ 3 1 ^ 6 4 1^0(161 ^1^0|3611^^

3.2 5110^5 1116

86160^1011 !!。^^. 1)1 1110 0330 0 『 3 000110110113 131*^61 ^31^31)16, 001 丫

1116 11~30|1131^ 1716(110^ 15 ^ 3 ! 1 & 1 1 1 6 0 0 0 1 1 0 0 0 0 5 , 11 56160(8 1116 11-3(111&1*6 0161110(1

^31131)16 86160(100 00(16 ^6^6015 111311^16 1 3 ^ 6 1 15

81


?\^\0\\\2 \40^1111^

11';7/7 848

&1161-^1-1^

^!!^!': ?1~001\00\ ^""(?,"恤8118111658

^1)11^0(10115

0:3(313乂 3.2

^9106

1 巾 — 6 0 0313 0加08(3 1*13X01355 1-6^81

100 50 06(31111

731091閘0(161

1.65

…―…….,—:罾…―::;

50 !'^)!咖01 ?355 ^ 咖 1 謂 ^ ! 0^3-3^1^^

!'X咖輕^3^1|3|9 卜

1

^ 輕 " ^ ⑩

1

^

"!^^^ 1^6!^"1005

3000 6.05 0.0050 ^55 丫 65 , 05^31111

^ 1-11^6 0(0^63[!0^1 !丄351曰^ !丄351 313^5

0011101918

! ^86(4& 10 6110 「丁1巾6 。「匕351 8"门 :

, 國 ^^^^^^^^^^^

11/21/06 1 1 : 3 5 ^ 111)5^1401-5^1-43130-91)61-18

40 11/21/06 1:59 0 4『,0誕23,20 8 6 ^

,0北^031

1)1 1116 ^1*0^611165 ^&!!^!, I 113乂6 561 1116 1 1 8 0 八 0 \ ^ 1 6 V;! 1*1:11)108 ^0^01^ ^ ^ ^ ^ ) ^

10 \^65 311^ 1116 1)56 01*011^

|3「0|3611^ 1 0 \ ^ 8 , 八 5 3 「651||1, 1:116八0乂16 VII!"1361(35 3「6 01^6316^1 911^ ^3550(5 1 0 廿 化 0 6 X 1 ;

1 1 0 ^ , 311^ 311 01355 ^31*135165 3 1 ^ 「 0 叩 ^ 1 ( 出 ^ 1 6 ^ 1 3 0 0 1 1 叩 5 6 3 》 丁 卜 6 ^31*131)16 86160(1011 110(^0 001151(^61-5 1)0111 1116 01"!^11131 3 1 1 ^ 八 0 乂 I 6 ^ ^ ^ ^ ^ ^ ^ III 1116 1^0-5(6(3 501601 !011 ^^!'006555 1)111 0 1 1 ^ 1116 361601:6^八0乂16 ^ 1 ^ 1 ) 1 6 3 31*6 3381^116(1 1116 「01(3 。!^叩加&\\6 |33556^ 1 0 廿 1 6 0 6 X 1 1 1 0 ^ ^ 0 啤 1 1 1 3.1 5|10\^5 3 56011011 110」已7116 0111(3111: 8|10\^8 1116 ^「!3131(25 561(^1x1 III 3 1 6 ^


^006^

3: 1^01-101)16 56160(1017 0/7^ 7^^/0^011011

。乂 1^0^101)165

83

011卬III 3.1

7。【: 11^^133 1^^125 1^^239 入0716 117^133 11^^009 10^16 1 ^ ^ 0 1 0 117^008 么0716 117411007 117X11135 !17^232 裊0716 11^^011 117^134 1^^231 !1^233 入0716 !1^136 ^ 1 6 07^235 &0^16 117*10138 87^139 !17X^236 !1^241 , " 11^^137 氲0716 117^234 117^011 117^009

1 1 15 9 10 9 9 10 12 12 9 11 11 10 10 11 11 11 11 11 11 11 1 1

V": V": 741:

1 1 1 1 1 12 11 10 10 13 8 12 1 12 12 8 11 12 12 12 12 9 15 12 15 11 8 1

V": 入0716 裊0716 40716 ^ 1 6 人0716 ^0716 息0716 入0716 森0716 裊0^16 ^0716 ^0716 人0716 ^ 1 6 螽0716 10716

V。!::

117^198 87^199 117^196 11^200 1^^084 抓脚33 117^012 1171110X4 117^013 !1^049 1^^166 肝^169 1^^033 117X11176 117^11179 1^*11177 1^^180 ^11(167 117111X73 ^11(086 117^207 87^168 87^061 117^048 1^111103 !^11^063 !^^090 117X11086

0.214804 0.170772 0.169009 0.152164 0.141193 0.140946 0.139130 0.137466 0.137062 0.137062 0.135750 0,135647 0,13564? 0.131550 0*131550 0*131436 0.131436 0.129268 0,129268 0.128856 0.127621 0.127621 0.125407 0.125106

0.082589 0.082346 0.082067 0.080211 0.076044 0-073751 0.067226 0.066869 0.064891 0.064249 0.061101 0.057662 0.057612 0.057351 0.056720 0.051881 0.05X642 0.050377 0*049710 0.048313 0.048313 0*048211 0.047419 0.047174 0*046800 0.046161 0.045998 0.045849

112 ^ 0110^2 112 ^ 11111112 112 ^ 010^2 1^2 〈 010^2 112 ^ 1111012 112 ^ 1110^2 112 ^ 11111^2 112 〈 3151^2 & 2 〈 11111^2 &2 〈 11111^2

0

9^ ^

04^040^

八111111166『0『^31*131)163 1 1 ^ 0 111611^16 56160(1011 (^!^^ 1》丁卜056 316|3 2 0『1116 \^18\3\6

X

^8(11131*6 00161^1011 111出6 ^61|111111317 ^31-131)16

\\^\& 3 0 11-5(^1131*6 6」0∼1116 111111111111111 1111*65^101(1 31*6 0 0 1 1110111(16(1 !0

56160(1011 口『00655,

1)1 316)3 2 , &11 1110 ^31-131)165

31*6 36)60(6(1 1|1 816|5 1 31*6 3116)60(6(1 10 & 86^11611(131『01^31x1

1*651*6551011 56160(1011 ^31*00655, 35 31501155601 66『0「6, 1|1 1116 560011(1 516|3, 0111 乂 3 !^!!爿尸"! 0^^31*131)165 111611116 8^0》1131*6

0!^6「;011, 35 0311 1)6 86611 &0111 0111|3111 3.2^ 丁 1115 00中111 0011(31115 3 1^516

13^60『『001 1116 0111(3111 ^ 1 0 ( 1 0 ^

1116 1^68111(8 ^ 1 0 ( 1 0 ^ 0『1116 \^311^16

86160(1011 00016.


?1-6^01^

1^10^1!^

1177/7 8^8

&1!,'-1化^11161:~ ^00(1001

80(1/(10115/。厂81(51)1635

/!/^/1001101^

011(^141 3.2

731115

5

口 ̶7511^

^214804 ^362606

^008124

^024565 辦 1 6 : !17^15049 匚

^7^239

3

^

3

1.145603

2.141463

1,393734

1.104802

678241

1.048243

^012379

.700517

222481

1.023194

^006199

^572663

^603923

1.000954

丁 116 ^ 3 1 1 ^ 1 ) 1 6 5 ∼ \ ^ 八 ! ^ 1 3 3 ^\\6 1 ^ ^ / ^ 1 1 2 3 9 3「6 561^010^1 1)01卜111 6111116(1

0 『 0 1 1 1 1 31X1 10 111611^

八II 01116!' 00111111110113 I叩"13讲6!"6 5 6 1 6 0 1 6 ^ 111 0111 乂 &61!" 1 ) 1 1 1 1 1 6 ^ ^ 1 1 1 1 0 1 1 ^ 9

5 1 ^ 3 1 11111111^「 0 ^ ^ ^ 1 3 ^ 1 0 5 \ \ ^ 「 6 561^016(1 111 1116 |3!"^11111111317 ^ ^ 1 3 5 1 6 5 355658111^111 (^^^

1夂111051 0 『

1116111 6\6 1101 1110611110 8 1 0 ^ 1^-8(11131*0 0^|1^1011 1 1 5 ^ !11 31^(3 2,

丁 116 110X1 110(^6 111 1110 ^ ^ 0 0 0 5 5 110\^ 15 1116 1^681*^551011 110(^0, 1 1 ^ 1 ^ 1 1 1 ~ 0 1 1 0 ^ 011 !1 311^1 5616010(1 1)1X11110 5 0 11131 1116

15 111x1316(1 XII15 11163115 11131 1110 ^ 0 1 1 ^ 5 1 0 5 3 6 1 6 0 1 6 ^ ^

8^10^(1011 1 1 0 ^ 3「6 ?10\^ |33550^1 1 0 1卜0 ^ ^ ^ 1 " 0 5 5 ' 0 1 1 ^ 0 ^ 0 \^11;11 0110^

011 口

, \^111011 15 !00310^1 1 0 1116

1110

「016 0〖111|1111 丁 0

1卜!5, I

0 ^ 1116 ^31*1^1)168 ^ 1 0 ^ 6 1 1 ^ 0『II化^0^1*8881011 0 0 ^ ^

丁1化\^31^1^13165 XVI 1 1 ^ 1 0 ^ 0 1 1 ^ 3 5 5110^11 111 0 ! 5 卩 3 3 ,

015|3137 3,3

口36 40716」輕8049 06^31111 严0^16—1^8133 06^811)1

師6

1^,31

了,

0'。& 「

"(!

叫 」

叩III ^1^*01國―」

^謂1 前!:舰1!^ ^^^^001 1^6002 1^6003 ^^^8004

^8011 !)^!]!! ~06^&110 ^6(81111 0^811)1

140 110 ^0 ―。 |40

[?61=0(6(1 [!6)6^6(1

^ , 1 ^^91

」 」

0811061

八【1116 10(3 0 「 11115 XVI!!^!^^,乜16 5 6 1 6 0 1 6 ( 1 八 0 乂 1 6 ^ 3 ^ 1 3 5 1 6 5 3「6 1 1 ^ 6 ( 1 ! ^ , " "

11化0011111111议0,6,

1101106 1)131 1)16 乂 3 「 ^ 1 6 5 3「6 ^ 6 1 1 1116 「016 ( ^ ! ! — . 丁 0 5 6 6 1116 1^016 3531^1121^ 1 0 1116

&5165

1 ^ 乂 八 1 1 1 3 3 3 1 1 ^ 1 ^ \ ^ 八 1 1 2 3 9 , 50!"011 30\乂】1 1116 ^31-1^1)108 \ ^ 1 1 1 ^ 0 ^ 八 5 ^011 50!'01 !, ^ 乂 八 1 1 1 3 3 560011165 ^ 5 ^ 6 3 5 5 1 1 0 糊 1 1 1 乂 | 5 ! 1 ^ , 3 5 5110\^11 111 0 1 5 | 5 1 3 ^ 3 丄

3丄11(3011 501-011111^ ^0\^11 1\|111,6『,∼\^八1^239

60000103


〔'^声V

101111011011

。【-/"/,/(?.、

85

0 1 ̶ 3 乂 3.4 、1 !^『1011166 -

^

0

&, 尺^^1 厂妹劇

一 ^ "

0|5|3|3^ 3.5

卜^『;2

| 1^50

!"'站^

| ^6

|吋 ∼ "

1

0^

^

1|

^ ^化"

^^?235~ ―

^^237――

1^31!"

061驢 !" 續 061續 ^|

&6^ 1^46^31

乜一

10 ^

卩―

110 10 10

"抓"

1^(6^01 」 ^ "^一

― 一

―1

1 ^ &^。…

^1-011, 015)31&乂3 3.4 311(1 3.5, II 13 016&!~出31 ^ 6 11010 ^ ^ ^ ] ^ ^ , 1^6031156 I

3.1

!"!31^5 | ^ 0 ^ 1 1 ^ 1 0

5110^55八0乂

1

^乂八!^1

只-841113「6 311^1 810(1 1^-8(11131*0

3.2.2

0*1001

136165 1 ^ 乂 八 ^ 1 3 3 &1\6 ! ^ ^ ^ 2 3 9 31-6 ^ ^

1^6513116【卜6 1^01 11131 0 1 6 ^ \^01~6 5 6 1 6 0 ( 6 ^

5 6 : 1116 1 / 8 6 八 0 乂 1 6

^ 0 1 6 3 ( 5 0 1113^ 3 8 0111^111 1116

1 01^

33

1116

1*131)16 80160(1011

6

" 廿 化 丁 1】!5 15

111 II化乂3!"1^16 8610011011 八 0 乂 1 6一1^乂八11239 ! 1 ^ 6 11101 1)0出

31101 1101106 3「6 ^ ^ 6 1 1 1116 「0)6 01"!II 1)111 ||1 11化

00111111110115 丁 3 1 * 9 6 1 ^ 1 1 1 1 1^011111131 0 3 ( 6 9 0 1 * 1 0 3 1

1 | 1 | 3 1 | 1 3 《 0 3 5 6 2\

丁 1化131-^61 ^31-13^16 1161-6 15 1116 531116 3 5 111 0 3 5 0 I, 3111^ 3)1 1116 111131115 3 1 0 110111111&1-50^16^

III 1116 0 3 5 6 0^11011111131-5031001 ^ ^ ^ ^ 0 1 " ' 0 3 ' 1111^1115 XV!111 3 0011111111005 131~^2^ 1^-5^ 113!'6 15 0310111316^ 115111^ 0 1 1 6 - ^ 3 ^ 八 ∼ 0 乂 八 . ^ 0 1 1 乂3「1311161 I ^ 6 0 1 ^ 6 ( 1 1 0 1136

化 6 0(511011 0『118111^ 6^!16「【!16 0 「 & " 1 3 ! 0「乜16 ^ 1 " 0 1 1 ? 6 ^

^「011 口6(1 0)355 乂3「131?165, 15 3 乂3「;313|6 03111101: ^6经!"011口601, 11 \^!11

51111 1)6 1110111^6^ 111 1116 561601)011 |31 1 00653 1 ^ 6 1 1 0 6 5 I

5 6 1 1 1 化 1 ) 8 6 01*01115 \^31135168 口「006(1^ 1 0

\^68 111【116 ^10(36(1165 |331161 ( ^ ! ^ ^ ^ 3 ^ 6 丫 I 1113(^6 11115 0110106 1^6031156 501116 0 卩 ^ 3 1 - 1 3 1 ) 1 0 5 1 0 0 11131!)' 0 3 1 ^ 3 0 「 & ^ 丁 1化161'!11 ^01(^1

\^

101)16^ !'6^6!'5 1 0 111050 ^3!'131)165 \\ 7 |1056 0316^01^165

-

( ^ 乂 已 ^ ) 3 1 6 ^113(^56(3 01 。。!!!匕!!^^!.卩0「 6X3111 口 ! " 511|3|5056 III61*6 !5 3 0 3 1 0 ^ 0 1 ~ 1 ^ 1 (!!^!!!!!!^!)

丁卜6 5 八 5 1111111111115 0 !'1011 1156 1116 101711 " ^ ^ . ^ " !11516311 0!""0310^01^;11-" I 1180 111650 1\^0 120115 5^11011^ 1110113 !)'


0 3 1 1 6 ^ 匸 1 卩 6 5 丁 丫 1 5 , XVII1011 111^103165 1110 1 1 ^ 5 1 ^ 1 6 0「"化01151011101-. II 1113^ 1 3 ^ 6 011 ^311165 511011 3 5 ^II 1-^311 0、^116「,,,

01-0!^!! 丁 「 3 化 1 饤 , , , 化 乂 3 「 1 ^ 1 6

031680「;65, 11 0311 1?6 00113卩5601 10

V

0!" 0316〇0「163

丁 丫 1 5 1135 1 0 0 1 ^ ^ 1 5 0「 5611111^ 1116 ^ ^ ^ ^ ) V I I 1*131)168

0「0|3611^ 10 \^08^

7 1 1 6 016501~1|3110115 0卩501116 0『1116 11011111131-50316(1 0 ^ 5 5 111^1115 111 !!))/ 6X31111316 ^ 3 1 3 561 3「6 ^ ^ 6 1 1 13610"':

0\^1^14

3 1:115101116「 5081116,11, 53561^ 011 I I 化 0 「 【 1 化 0 1 1 5 ( 0 1 1 1 6 ^

^3)116 15 6 3 5 6 ^ 011

& 0 1 0 『 5 811011 3 3 11110 !'651 1 ^ 6 0 0 6 @6116!'316(1 \ ^ !116 0 0 5 1 0 0 1 6 1 \ 71115 ( & ^ 6 3 0 0 ! ) ' 1\^0 乂3!1165: 9 ! ^ 1 0!" 1 0 仏 0\^^\1103

3 5681116111, 1 ) 3 5 6 ^ 011 1)16 1^(36 0『3000111115

0\^11!3

3 ^31~!31310 1|化1(^11118 ^11611161 1110 01151011161 11565 1116 111160161

0\^^^07

3 霍 1 & 1 ) 1 6 111^10^1:111@

。\^八1106

3 ^31-13^16 111^1031111^

-

1

6 ^ 1卜6 01151011161

-

1

-

[!16!' 1116 01131011121 0011130(3 6乂 ^ 0 0 6 1(16 0115101110!-

出 6 131*311011 ^6(^1101111^

XV 116(1161- 1!化01151011121~ 1 1 0 ^ 5 1111111^16 3000111115

3 ^31-13^16

0 1 5 口 13乂 3 . 6

:117)150116(1 0313 [-^^^0^01 03(3 ^!'1313135 ! |^3!? 0 ) 3 5 5 16\1@1

100 50 08131111 ^63 ^83

! 卜 他 " ^ 0(81113 \ 1-^3X111111^∼"&!^!^ " 馳 咖 0 1 0卜卜5叫3『6

50 6 3.84

1̶!:幽凶腿01乂?化^."謂一 1-1^111111(111111 ( ^ 科 円 ! 「3【0|3 1^-8^1131-8

3000 0.0050 15.05-4

! 1-1036 1^6^16 ^3!'13&163 ^ 3 6 0「0叩乂3『1313|65 11118 !'301100 3 ;:8?03 ::^||110011011

1^0 丫6 5

\ 1-111116 0^ 0^631100 1-^110 1。 ;::-;匕351曰170『 ! 1-1-331 3131115

11/21/06 4:52 ^ 0 。11357159-3668-4343-822-3

8^0; 1 . 3 5 1 編 ;1-^11(1 011 !'311011 ;;:6「!。 1-1051

^63

00一

6!6

11 ^21105 4:59 ^^1 0 !")「 0 ^[[!. 15.20 3 6 ^

0 1 1 ^ 1 1 1 3.3 5!10化5 1116 ^31^135165 5 0 1 6 0 1 6 ^ 111

1 0『1116

8616011011 1 1 0 ^ ^


^017(6?" 3: 1^01-101)16 8&1^011011 0^

!)^"^/。"議II。》0^ 1^01-10^168 8 7

0111(31113.3 ' ― … " - " : ― …

「',.'-一"―-―:…

國 … - ' ― - ' ' - . ―

~:~~圍

^、、-

:1116 1)1111^ ?1006^12X6 1^-3(11101:03 ^01: 丁 。 7 6 X 1 6 1 ) 1 6 : 。!:?乂

01&33 01&33

0^14 0^03 07^03

1 10 4 2 1 2 1 2 1 2 4 5 4 4 3 2

01&33 0^0X1&

071113

01&33 !?!:。"!)

01&33 6101^ 01&33 01&33 01&83 。【。!!。

01&39 &"叫

01&89

0^07

07*10.6 0^05 0^1105 0^08 071108 07^10

0 173336 0 048790 0 048027 0 020582 0 020577 0 014605 0 014560 0 014177 0 013946 0 007594 0 004338 0 004122 0 004120 0 003097 0 003067 0.000041720

1^2 112 1^2 1^2 1^2

1111^2 111^2 0119^2 111(0^2 1110^2 11113112

丁。 566 1 1 0 ^ 60161^1*156 ^ 0 6 『 @ 『 0 叩 5 0『00113(3363 03(6^01*1031

&59 00115!016『1116 11011111131-

30316^ ^81*131)16 0 乂 八 1 1 0 3 , 8乂 1131113 ?1^00 ?^.^ 1(1 1116 & 4 8 0 0 ( 1 6 00(16, ^ 6 0311 366 111311116 0 乂 八 尺 0 3 1135 1 1 0 & 1 6 8 0 1 ^ 6 5 , ! ^ ! ^ ! ^

11011111131-50310(1

1111331113 0『1111131)616^ 0316^017;

85 5110^0 111 00中111 3,《

0111 口 1113.4 7116 11^0 ?1:006^111:6 011111118&1^6 ?1:0(1116110^

01 02 ;03 04 05 06 07 08 09 10

101 52 60 126 228 156 148 130 104 54 9

?6I:^6]1^ 8.65 4.45 5.14 10.79 19.52 13.36 12.67 11.13 8.90 4.62 0.77

101 153 213 339 567 723 871 1001 1105 1159 1168

8.65 13.10 18.24 29.02 40.54 61.90 74.57 85.70 94.61 99.23 100.00

丁 116 ^31*131)16 86160(1011 00(^6 01*6316(1 & 0 6 ^ ^31^31)16

0 一 0 乂 八 1 1 0 3 6 ^ 00113|351113 1116 1 1

0316呂0「165 0『0\^^\1103 1)110 5 0316呂0「165, 115102 ?1100 ? ! ^ ^

^ 6 0311 366 1卜6 ^ 6 ^ 1 1 6 0 0 ^

^1511-11)1111011 0『1116 0^6801*165 0 ^ 0 一 0 \ ^ 八 1 1 0 3 , ^111011 13 3110\^11 111 011^111 3.5, 「01116 5 八 5 000*0 0&11613.1&6 011(131118 3.4 &1\6 3.5, 5 6 6 015口13乂 3.29 !0 1116 3(3^)611(1 IX 10

0113口 161\


3 卩 " 8 3 53!」033183 0 1 ^ ] ^ ) 133,181

53111(^11103 3」II口330』3 3 1 卩 ~ 0 1 ~ 0 111(111!

1133^43^ 3 纖 ^ 习

3 3 ^ 0 3 3 1 6 0 31(1 5911!^11103 ^」^;);)。」!:' 3 1 ( 1 丄 9 51111 3 5 0 ^ 1 1 5

3

0 '!3 ' ; 3 :331-19111335

0111^)113353^ III ^ 3 ) ^ 0 3 1131(1 3 」 6 5 3 卜 1 0 3 3 1 8 0 3 1 | 丄 习 ^

2 』 0 3560 4

0

^01 31卩1110』』3111^3 III (^)!"^」311!」3|3」0

110 0 5 ^116 1 X 3 1 1 5311100 1163111 133」巧1531)3(11 1X911 31卩1!1!八、0110

5 3 1 1 1 0 0 3 ^ ; 3 ^ 6 1 31(1』。』1163111 1531|3|1| 31(1 561| 161(1 ; 0 0 8 9 1 1 3 3

^1 ' 卩 八 3 | 岡 1

5! 34611 !35"^! 3 1 卩 0 5 0 ^ 1 1 5 "(^!!!011043

屮 3 9 岡 1 5 4 ; ^

^ 0 3 3 1 0 0 15.1!^ 3 1 卩

'?; 181!丄,51163111 133.151』!31|1

35110(^53.1-11011

I

1

13^10

351100153』111353」8^ ;

1111|1;^\ 5^3(3110(139^ ^ 0 11011.10(10^(1 31卩51 |3八3| 113八!8 8 』 0 ; 1123111

31(1 1^181义11,11!0| 4

1!| ^ 3 1 6 | 1 1 0 | 8 3 8; !3八3| 1 叫 1 III 5口」033」31|1 .10』193^6; 3 1 | 1 』 0 111331;! 31(1 | 3 八 3 | 1(069 ^ 0 ^ V

: 3 3 ^ 0 3 3 ^ 0 3 八 " 3 1 0 ^ ! 』 8 八 5 5 8 | 0 8 9300101118 " ^ [ ( ! 111^X3 0|0111115 8 ^

|531^11511||1 3 ^

1180 3闪8!」13八558【0 8 』 0 501^0^9)133 31(1 811|30151|03 II! 5 3 5 1 1 」 3 1 1 ; ^ 35|^^13;113 3.111(3300.1^

^110.13 31!丄

一卜 ^1 、 : 一 , " —

0 : 0画 0星 0 ^ I ^

6-2 22

^,.)^^

3 I

』&,'

7.1

^ 單

\1

3 1 ,,^,)。, 9 1 9 1 1.5&") 1 ? " ) " , 0

; ^

1\

!.^^)^^

9 I 1. 二 I重

02 67 81 91 57 6-1 27

3 1 9 1 ,, II 5 1191111 "卿貝一化叩 ^

, 2腳"一 门^

^ ^ 3 1 1^11

,―卿&―

"叩晚

,

1^

9917

55*56

9111

29-6^

026 269 9^

51?'^

^5|018!0

2? 981

【~0

00*001

^31^ 邦3

00*22

1X1

1

1^1

I

9^

0

0旭3111

99

10^100


3: 1^01-101)10 8616011011

0\00^^

311」1-603101113165 1116化―5只113「^ 1 ^ 0 ^ 1116

0)1^乃-"〃々'/^〃?'"〃0〃

。/^"〃/"/,/(:'.、'

89

1135 10111~ 。 3 ( 6 ^ 0 「 & 5 , 311^1 1110 1^-5(^1131^ 1 ) 6 ( ^ ^ ^ 1 1 出 0 0

111^111 311(1 1110 1 3 ^ ^ 1 1 1 0 ^ 1 ^ 1655 111811 0.10. 1 「 " 化 [ - ^ 1 1 0 1 1 0 1 1 111 1^-8^1131-6 13 1^55 111311 5 ^1, 111611 0 311(5 5 1-0013111 0011113! 116(1 ^八『6(^1!"1011 0 卩 1 1 1 3 1 1 5 ^ 0 111 311 116^

111卩111 1111151 ^

6

51^13「6

0厂0.20

11163115 11131 1110

3 0 ^ 8 ^ 1 1 & 0 「 31 16381 0 . 1 9 \^!:11 1116 13 !"^^. 1「 1101, 0 3 1 1 ^ 6 3「6 0 0 1

001111)1116^ ^ 711611 311 3 1 1 0 1 1 1 ^ 15 1113^6 10 00111131116 5 311^ 0 115111^ 11115 531116 0161110(1, 30(1 5 0 011. 0111卩III 3.3 3 1 1 0 讲 3 1110 1^-8^1131-6 1 ) 0 1 ^ 6 6 1 1 出 6 131^61 311(1出6 01"!^11131 111|31|15 1 x ^ 0 1 ^ ^011(31113 (!&15016^ " ( ^ 娜 , , ) 3 5 \^611 3 5 1116 11-8(^113^6 1)61^0011 1110 1 3 1 ^ 6 1 311^

^ 131)616(1

;;

8『0"!36爿乂6「5!011 0『11)6 10)31115

0「0叩,,》

0.048790 、 " 化 【 1 化 1 3 ^ 6 1 1^6^01*2 115 0^0.048027 3&。(^、'。「(! 801(1 (^―0\^八1《03

丁 116 11011111131-5031^ ^ 3 ^ 1 3 6 1 6 〔 V 八 & 0 3 1化5 311 1^-3^1131-6 0 「 0316^01-165 3「6 001111)1112(1 0「 3「011156|1 811^ 011 1^~8^1131~6

&11(1 0\^八1103 17101 1)16 1^.-3^ 1131-6 0「&6「!011 311(1 \、《「6 56160(6(1 ! " 316)5 I , 9 。 \ \ ^ 6 『 , 0 0 1 ^ 1卜6 ^1~011|104! V!11-131)16 15

111 1110 II 11*1! 3 1 6 ^ 2 5 6 1 0 ( ^ 0 1 1 , 3 5 0311 1)6 56611 III 01)1(3111 3國6, 丁 116 ^~3^1131~0 5110\^11 111

01|1|3111 3 . 6 !5 311 '"化「2111611131 , 0「 ' ^ ^ ! ! ^ ! ! ^ ^ ! ' 1^-3^ 113^0 3 3 ^!5111101『「0111 1110 61乂3「;316尺―5^1【13「6 ( ( ^ 「 卩 ^ ! ! ! ; 0 ^ ( ^ 加 " ^ ! ! 。 ! 】 ) 5 1 1 0 ^ 1 1 111 0111)0111 3.3,

0吻111

3.6

71^ 0111111:卩!:0"(!!!" ^01: 丁3,匚

8 - ?!!^" 1 4 1 1 1

^^03 01: 0X111: 0^13

0.173336 0.009016 0.008175 0-006723 0.003592

^化1化

5^

2 44.497591 3.203197 11,725295 9,715610 5,210083

《0001 0.0125 0,0006 0,0019 0,0226

544.655917 28.329505 25.637918 21.126479 11.283295

2.227745 2.211033 2^190812 2^174488 2^166625

丁 1化111-51: 、3『;313|6, 〔 V 八 ! ^ 14, 15 3 ^ 0 - 1 6 ^ 6 1 0 1 3 5 5 乂 ^ ! ^ ^ ! ) ^ , 5111^ 11611^2 115 1 ^ 6 1 5 3「6 1101 0 0 1 1 3 ^ ( 1 ( 狄 0 1 1 1 ( 5 1 1 1 3.3 311(1 11010 :1131 110 0 「 0 1 屮 乂 1 ^ 1 0 1 1

0化\^14

3 叩 6 3 「 5 》 八 5 1^「0「6, 1116 0(131 5610011011

15 1 3 3 3 0 ^ 011 1116 5 1 6 ( 3 ^ 1 5 2 「0「\丫3^ 3610011011卩「。(;^!山力.丁 1化卩「0060111「6 513「15 XV 1(11 1116 13651 " " ― " 山 1116 ^ 1 * ^ 1 6 讲 油 出 6

^ 1 1 ^ 6 5 1 0 0 0 * 6 1 3 ^ 0 0 , 「011046(1 !^' 1^0 0【1化「 V&)'I&5)63 3 0 0 0 ^ 1 1 1 @ 1 0 III6!!"

001111"!1)1111011 1 0 1116 !11,13「0^61116111; 0 「 1116 510|5\\;156

11-8(1113

1^6X1 1116 1^6^1*6881011 1 1 0 ^ 6 15 0 0 0 1 1 6 0 1 6 ^ 1 0 1110 ^31-131)16 36100(1011 1 1 0 ^ ^ ;111(1 1)16 乂3「1313|65 !:1131 3「6 ^ 3 5 5 0 ( 1 0 0 【 0 1116 ^ 0 ^ ^ 0 8 8 1 0 1 1 110(^6 XV! 111 II化1^010 0「1|1

013(3137 3.8

^110 06?^

| 1^ ,

0 0^^07 06131贞 0~0\^13 0 6 續

| 8 一 加 ~0 ?10

| ^010 帥III

\

1,1 ,叩口 1 1响1 10卩^

~閥1口31

10 》

1

^01711^131 ^01711^181

」 ^^。,

|「 01^ |

|

|

3.8 5110\\^


90

?1-6(11011^6油^/"兮XVI'出8/18 ^6^186她6广6^11(^1

3.2.3

8118111688爿/ ??/ //"//衝

80^0)15/。?"

6 1 1 1 3 1 ^ 1 * 3 1 * 9 6 1 ^ 8 1 1 1 11111116110 1 1 1 ( 6 1 ^ 3 1 - 8 0 3 1 6 1 1 誦 叩 " 1 8 《 。 3 5 6 3 》

10 1(113 6X8011)16, 3 ^^01116(10^1 1)31^ \^3015 10 01-035-8611 3 010118&8 『6003006 " 。 ( ! " " . 丁卜6 5 3 0 ^ 6

560(15 111311 10 1(3 6X15(1113 0115100161^^ 丁 116 X&^1 0『1『0『『65530056

\^\^\^

13 1^68|301156 10 1116 013!1, 13^108 ^ 6

0 ?61 000-1*651)01156^ -

7116『。110^08 13 3 3 6 5 0 1 ^ 0 0 0『30016 0『1116 1叩1115 III出6

I& 361 1136(1 10 ^ 5 6X3011316:

031X1 5^131106

^ 八 議 4

311(0 10311 6&1&1106 311(0 16356 1^八11030

動 ! ^ 3

^八11044

6

10&11 1?&1&1106

10311 0 ^ 01001111^ 1)^11530(10118 III 01160^1113 30(1 33^11133 30001111(8 00111156『0『6^01附016015 ^

1卜6 63(1^ 口6『01011111

3^31181)16 01*6(111: !!"!!!!^!1 0『11^65101601-1*61316(1 (^(!^^^^!!^ |36『0100111 1)61*0611(336 0^11011361101(13

01111(11*611 \11\6^^ & @ 6 10 111出6 1)100^

2X0^

1^八11195

。61*0611(336 0^11011561101(15 1卜!11@ 111 1*611(6^ 1100165 111 1116 ( ^ 。 ( ^ @『011 口

~ V八!1210

\ \ \

0『1*6310161106

& 8 6 0『1116 011510016『

^八11318

111101

0『131*0(1110(5 0^06(1 1?乂 1116 0 0 5 1 0 膨 「

!!賺!)"

0『16116^1^053011003

V 111011(11

10 1116 0356 0^1)111317 I & ! " 8 6 乜 , 出 0 ^ 1 1 1)6 3616016(1 ^

1151(1^ & 1 1 1 0 11-8(11131^6 0『116『100, 33 111

0 3 5 6 1, 01-1116 0111~8^1131^6 01^116^0114

3.2.3.1 8-〇0|113『6 01-1161-1011 \^116111116 11-3^1131-6 0『116『!00 !5 1156(1, 1卜6 乂3!"!31)16 56160(1011 13 510111&『10 1116 0836 111 ^ 1 0 1 1 1 1 1 6 131*861 13 000110110113 ^0&5&

1, &1^0V^)^ 丁卜6 1 3 ^ 6 1 ^31*131)16 13 11*63化3 111(6 21 000110110115 乂3『|3616, &1\6

1116 ^ - ^ ^ ^ ^ ^1111 1116 131*^61: 13 0001151116(1『0『63011 0^!@11131

&1\6『0『63011 1)111116(1

( 八 0 \ ^ 1 6 \3.V\8)0\6.) 丁!160 1116 115031 1^0~816|5 131*006(1111-6 15 ^0110^6(1: 111 1)16 ^11^1: 516口, 3 0 1111(131 ^31*131)16 361 13 36160(6(1

00

1\110111111111 ^~8^1181^6 00161^100; III 1116 560011(1 516口, 3

36^11011(131『0I*\V3『^1 56160(1011 131*006(1111*6 13 1136(1 10 561601: ^31*131)163 6*0111 111036 56160(6^ 111 316|3 1 0 0 出6

0 『 & ^ ^ ] ) 11-8^1131*6 0『116『1011, 111(311(8 ^111011 1116611)0111 111636 01*1(61*13 1*606^6 1卜6 1^016 0『

111(101 III 1116 3111)36^116(11: 1X10^611113【001, 31101\ 35 1116 ^6^1*6881011 110^6^


0 ^ 1 ^

3: 1^0^101)16 86160(10^1 0^

丁而3(0"賺"011 0^ 1^1(31)163 9 1

3.2.3.2 011~30|1181^6 0『:16『:011 XII13 01-1161*1011 0 3 0 6 6 1136^1 0 0 ^ ^ 6 0 1116 1 3 ^ 6 1 13

^ 1 5 01*1(61*1011 13 5616016(1,出6

1

36160(1011 |31 00633 ^ 0 6 5 0 0 1 113^6 1 ^ 0 0115(11101: 5(61)5, 38 111 1卜6 0836 0『1116 11~8|| 1181*6 00161-100 ^ 1115163(1, 3 11-66 15 00113(1*110(6(1^ 丁卜6 111^11(5 56(60(6(1 III1卜6 00115(1110(1011 0 ^ 1116

&1*6 口&8563 10 1116

06X1: 1X06& ^1111 1116 3 5 5 ^ 0 6 ( 1 ^ 0 ( 6 (^】!!!)!^. 1|1 11118 6X8011)16, 11161*6 31~6 3 0 2 10|5018, 311(1 1116 86160(100 110(^6 5616010(1 1 8 10口1115, 丁出5 3616011011 13 ^0116 6乂

\0^\1\^ 3 11*66 1)35601 0 0

1116 0111-8^1131-6 0『!16『100, 3 3 6X^)13106(1 !?引。^.八加6 00113(1110(601 1131113

0111-5(^0&!"6 0『116「!00 13

500161111163 08116(1 0 1 4 八 1 0 (^!!!-^^!!^^八010013110 111(61*^11011 061601;00乂 丁卜6 11-60 3 6 乂 6 | 0 ^ 1 1 6 0 1 1^0066(15 38 ^6110\^5^ ?!^!,

『6001X15 0 ^ 1)10 6&13. 561 31*6 ( ^ ^ ! ^ 111(0 1 ^ 0 1

011

55

31-011135. 53011 31*01113 ! 3 ( ! I V 1^6(1 1010 1\^0 0101*6径 " ! ,

5 0 011. 丁卜15 ^1*00653 0 ^ ( ! ^ ! ^ ! ! ! ^ !8

03116(11^6014^81^6 / ? ^ / / / / 》 " / " ^ &30(1 31*011(3 01*63(6(1 ""『比^ 1卜6 1*60111-31^6 133111(10111113 13 03116^ 3 " 。 ( ^ . 丁116 110(165 11131 31*6 0 0 1 (!!^!^^^ 1\111|16「 31*6 0^116(1 16^11101110^689 出 6 11-66. 丁116 110(165 11131 31*0 ^1^1(16(1『011116『3『6 03116(1 1^6^6(11016 『6001^5 0『1116 63.1& 361 13 0^116(11116 ^001 "(^.八II

0 『 7 2 0 ^ 5 , 0『16(^65 0卩

^0^65^ 7116 110(16 ^1111 &11 1\\&

111636 110(163『01*01 1116 XX&&. 0 1 5 口 1 ^ 3.8八5110^5

3110(1 & 10 01^6『10 5|3111;出6 1^600^5 III 3 ^ 3 0 2 111 11113 6X3111 口 16〉 0116

3

110(16 111(0 1 ^ 0 0 6 \ ^ 110(165, 016 31801*1111111 6^31113165 &11 1111)11(5 &, 31)6 1111(15 1)16 1)651 5(3111『0『630)1

『0!" 3 0 10161^31 10|301, 1卜6 &1801^11111厂11*51 ^ ^ ! ^

丁0

1116 1)631 5|3111

1116 111^111; 111(0 10161^315? 0『1)105, 0^6^1131 160^11 ^

丁卜6 !!"!!!!^!1 0 ^ 61113 111(0 ^ 1 0 1 1 1116 111(61^31 !0|3111 15 ( ! ^ ^ ^ 15 ^6(611111116(1 6 丫 化 6 ^^1116 10 ^111011 出 6 ^ ! ! ! ! ! ! ^ ! 1 0『8!08 口『0|3611^ 13 361. 丁 116 ^16『311|1 乂31116 13 5 0 , !~10\^6乂6『,511131)056出6 1156『36(8 1110 10 31\6 111311116 10161^31 I叩11110 1)6 1)11111601 13 10001116 " 出 ^ 1 1 1 6 5

!'311^6 ^ 0 0 1 3 1 0 , 0 0 0 10

8 1 1 0 , 0 0 0 !01116 6 ^ 3 6 1 . 1 0 I I I & 1 0 8 8 6 出 6 1 0 6(1113110^1^315 \^;111)6 5 1 0 , 0 0 0 ~ 5 2 0 , 0 0 0 , 5 2 0 , 0 0 0 ~ 5 3 0 , 0 0 0 , 5 3 0 , 0 0 0 ~ 5 4 0 , 0 0 0 ∼ 8 0 ( 1 5 1 0 0 , 0 0 0 - 5 1 1 0 , 0 0 0 ^ 八 ∼ 1 1 0 ^ 6 3^1 !1 ^8&6 0 0 出 1 5 ^口01化1111)0 133111(1011601 &I 0 0 6 0 1 * 1 ) 1 6 卬 ^ 、311165 0 ( 5 2 0 , 0 0 0 , 5 3 0 , 0 0 0 , 3 4 0 , 0 0 0 ∼ 八 2 x 2 00011086110^ 15 01*6&16(1『0『030(1 0『111650 111116 5^)111 乂311165 &11(1 & 0(1卜8(11131*6 5(3(15110 15 001X1151110(1『0『63011 5卩1 10 11115 6X3111 口 16, 11161*6 31*6 111116 0 ^ 1 - 8 ( 1 II&!'6 乂31065, 7116 5|3111 乂3106 "!也1116 111@11631: 0111-3(11131*6 乂3|116 15 1(16 1)651: 5口1111)350(1 011 1\\6 111(3111: 10001116. 丁 116(1出6 3130『11|1111 ^ 0 6 5 3 5||11;13『03101113(1011 10 110(1 1116 1)631 31)111: 0 『 1 1 1 6 ∼ ( ^ 『 ( ^ 0 0 出 6 5&313 0 ^ 63011

0『1(16 011161- 3^311&516 ^!^^.八&6『^10(1108

1116 1)651 5^)111『0『63011 !111)11仁1116 31^01*1(11111 00111|33「65 1)16 0111-5(111虹6 乂311165 0『&11 111656 1)631 5卩1115 3.116 1)101(3 1卜6 0116 11131 @ ^

6 8

0 0 1116 1)3513 0『出&I !0|31111^18.1

1^^1651 0111-8(11121*6

II ^ 6 1 1 3^11(3 1116『^^『(^ 1010 1 ^ 0 00(165

出 6 1)631 5口1比丁 1^3 ^1-00655 15 1^6|56316^1 3.1 63011 00^6^八II 111056

1111311(3 11131 31~6 56160(60! 1)1出;3卩『00655, 1.6. 111050 1卜&I &1*6 01103611 10 1)6 1^16 1)3513 0 ^ 1116 3^111 31 311 乂 110(16 0『1116 IV&&, 31*6 「355601 1。 1^16 116X1 110(16 " 出 1 1 1 6 3381^116(1 " 。 ^ 0 『 1 叩 仏 7116 11-66 ^ 6 1 6 1 ^ ( 6 ^ ^

^

\^31^13516 86160(100 110(16『0『011『0X3171)316 6&1& !5 5 1 1 0 ^ 0 !11 0!5口13乂5

3.8八3001 3 . 8 8 , 0 1 5 ( 3 1 ^ 3.8八5110\^5 1116 1*001110(16 &0(11116 16^-11311(1 1131^0『1*16 XV^116

013^)1^

3 . 8 8 5110\^8出6 1-001110(16 30(11116 1*18111-1180(1 113^0『1116 1『66, 1)6031136 1^16 ^11016 1^66 13 1 0 0 ^1(16 10 ^11 III 0 0 6 (!̶&乂.


V 1^(1X01^

1^06^11^

"'〃/! 848丑"!,.!、化應化!': ?1-00(10611 80111!10!1^811X1)16^ /!^!!00110(1^

0 : 8 0 | 3 ^ 3.8八 1

015(3137

3.88

~ 3 ~

0 ― I ―

1

3

~ 0 ~ 瞧^2∼

瞧8075

0111^111 3.7 5110\^3 1116 1*61 & 1 ^ 6 1111(301131106 0『1116 52160161^ ^ ^ ! ^ ! ^ ^ .

13 隐??014


010陶-3:

^0)1(1^16 8&160(1011

丁,'011^0!"賺!!0(1

。〖V。"(:93

0111(31^1 3 . 7

円!6 逮

1: ^16^ ^10^01^ 崖

^ 柳55,731^1 匕麼匕

0

^ ^ 1 8 ^ 0「 ^ 1 ^ 5 1^ 丁[^已6

^61311^6 !^...

^^??315

1

键6137

附朋137

0.823072

似6210

0-648884 0.61899

^ ^ 1 2

^^212

0.606703 0.601633

则 , 4 8

0.600437

^ 屬 7 5

0.550595

^^^318

0.525406

∼八6286

0.56696

^ 鳩 1 6

0.527099

似6318

0.507789 ^^193

93

0.504904

權6275

0.501803

幅6187

^化187

0.471972

^化230

^ 緩 3 0

0.432374

幅[?5^ IV八^]04

0.429983 0.367988

^^^014

则^014

丁卜6 丫3「;36165 11131 丫00

0.343943

111 015|3!3^5 3.8八3113 3.88

0311 ^61 - !!^ 11118 ^

|33556^1 10 11化116X1 110(^6

1:116 1^,010 0 「

("!!'51 0011116011111: 3 ^ ^ ^ 1 " 6 5 8 ' 0 1 1 10(16 1 0 出 6 ^31*131)16 8^1601:1011 1101^6-

& 5 5|10\^11 111 015(31&)' 3.9^

0:3|313^ 3 . 9 丄

劇610. 10100^80 0313 也 ,

^':II闩1^:0 |^!^」「―^!?「【!1。。 5 「?^!: 1)7101,1^1 丁 9^『15 ^0^10^1^ 。&0'&琴 卜"56【10—5

「』1 1:个」1 [ ^ ; 0

"丁

:口0"飄咖

丁I化II 1『乂011 5 6 1 ^ 0 1 1116 ^ 6 ^ 1 - 6 8 5 1 0 1 1 110^1^ 1 1 ^ ^ 1 0 【 I 化 ^ 3 1 ) 1 3 1 1 ^ 01|0!\ 011 !^] 10 1116 「1^11 0 ^ 1110 口 『 0 口 6 ! ^ 111 1116 |310|301~1|65卩31161, 3 \^111^0^ \^!11 3 叩 6 訂 ^ 1 5 ^ 1 3 ^ 1 1 1 5 出 6 ! 叩 ^ ^ 8 5 5 6 ( 1 1 0 1;116 ^ 6 ^ ^ 6 5 8 1 0 1 1 110^1已丁!!!5 15 5卜0\^11 111 0 | 5 ^ 1 ^ 3.10.


^1-^1011^6 ^10(1&11)1^ ^1(115^5

^:咖\41獻:?1-00(100180111(10)15恤8!"!!而^!100110^

0;5|3|3^ 3 . 1 0

251 "叫

3

^ 卿 ! ] "

01^

0^911)1

00

^ ^ 1 4 8 06*31111

^ ^ ^ 2 1 0 06^1111

1^/^275 0^311^1 ^ ^ 2 8 6 06'讀 固 8

^6(3"!I

1\

0 1 5 ^ 1 ^ 3.10 5110\^5 111311116 1^010 3551^16(1 1 0 63011 0 ^ 111656 5616016(3

1^八111 87

311^

1^\^尺315 3!'6 111111,131^,

0『1116111 1135 1^\^12「

111111146 1:116 1)00181011

2 0 ^15111101

3 「 6 让 ^ " ^ 已 ! ! ^ ) ! ^

1^乂八111 8 7

1135

14 1^151:11101

'!3^165 15

1^0(6 11131

3 5 11011111131 511100 63011 乂3!1165, 31101

乂八113 1 5

1135 7

110(16, 1116 ^31.131)16 801^0(1011 110(^6 ^ 0 6 5 1101 01*63(6 1^11111111^ ^^!"!^^!^

&0111 1116 161-11111131 110(^65 1111(^6!' 1116 0111-3^1131*0 1;「116「1011 丁 1,6 0111-8^1131-6 0「;^『;011 15 ^ 1 5 0 1 1 5 5 6 ^ 111 11101*0 ^013(1 III 0113口16「 4 111 1116 001116X1 0 ^ ^60151011 11-665^

0 1 5 1 ) 1 3 ^ 3.1 1 5110\^5 1110 131*013611165 13(11161∼1化!^ 11化01~00611165『0!" 1116 0111-8011&1*0 ^!-!^!'!011 31*6 5 6 1 .


('/'"/?&'' 3:

"!8616011011

011(1 71-0)1^0)1110110^

0^ 「(?/.,'"/)/^'

95

0 1 3 ^ 1 3 7 3.11 ^^1116 ! ^0(^610 101(50116010313 0 化 3 03(3 乂3|13&165 」[!43X01355匕∼61 : 1 ^ 5 5 1 门 9 (^「(:^柳 ; 1 3 ^ 8 \ 1^0(^61 1 ^6|6:1 1^111^56^ 73!'13)3163 1卜"1101!16「 0(31[!5

^

^ ^ ―

100 50 丫65 ^68 50 6 3.64

1

早 ! ^ ^ ! ^ ! ^ ! ! ! ! ! ^ 3000 0.0050 506-4

;-^11111111,1171 ^^3^1^16 卜510口 ??~5^^3?6 1-1139 ^ 0 7 1 6^91-1315165 卜。56 0101^31131:165 : 力36 11116 ^30410115 51=05 ;?'1-1111 001100

^0 ^65 1^0 丫6 5

11/22/06 4:28 ? ^ 00424662-6800-460^910^9~

:'711116 0^01-6 311011 1 !丄35化110「 ! 1-1.381 313(115

0001^)1816 ^0 ^10 11/22/06 4:38 0 I"!「― 0 隐 . 3 8 . 7 8 5 6 ^

1

1-^1660)5 "10 6110 ― 陶 卜"6805 ! 71[!16 0^1.351 ;卜??00 001-3(100 : 0【1。1"!051

3.2.4

8 1 1 1 3 1 ^ 丁 3 1 ^ 6 1 ^ 1 1 1 1 ^ 0 1 1 1 1 1 1 3 1 - 5 0 3 1 6 ^ 0 3 1 6 0 0 1 ' ! 0 3 1 ||1|31113《03564 ^

|11 1(115 0 3 5 0 1110 1 3 1 ^ 1 ^ 3 1 ) 1 6 15 " 化 5 3 1 1 1 0 3 5 :11 0 3 5 6 3 , 311(1 311 1110 ;叩1115 3「6 11011111131-50316(1 0 3 1 6 @ 0 1 ~ 1 0 3 1 1^61^ 3「6 ^ 5 0 ! " 1^110115 0『801716 0 卩 ^ 1 6 11011111131-50316(1 0 1 3 5 5 ;0口 1115 III 1116 ^ 3 ( 3 861:

^ 八 1 1 0 8

5 1 2 6 00(^0

0\^^1^1 1

10001116 00(^0

匸\^八1120

^ ^ 。 ! "

〔V八&34

0『1101116 0\乂,16「5卜1卩

;11310310「0厂1|116「1化1 6311^111^

0\^八1135

0115101116!" 1 ^ ^ 6

"

11(6

0 \ ^ 4 1

0^016 0 0 ^ 6

\\\^\0^10^

0 ^ 111(61'!]01 1153^6

1-161-6 3 1 5 0 ^ 0 1 1 0311 1)50 6111化「出6 !1-3^ 1131*0 0「 1)16 0 1 1 1 - 5 ^ 1 1 3 ^ 01~1101~1011 1 0 501601【I化乂3「;313|65國


?)'6^10! ^

3.2.4.1

XVIII!

、、,1111

140(1&11^

848

3

5咖.[?,'!化!^!^!-: ! /^!!^! 8011/110^/0^ 8115111^

8-3(^1)3^0 01161*100

2 36501'!1^6(1 111 1^1^61 1X10(161 口!"0口611^ 1 0 1^

11115 01-1161'!011廿化5616011011 ^!'00001II 1^6 15 51111113!' 10 1116 |3^006^11^6 1|1 0 3 5 6

0:|13口 16「 2 (^&^1\0\\ 2.1 8 ^ 1111 !"0 30(5

1.0 111

XVII!011 1\^0 5(6(35 3 ^ I 化 ^ ( ^ ( ! . I 5^1 1110

1112 乂3!"!31)16 80160*1011 110^6^ 丁 1化\^1~!35165 5616016(1 1)1 51613 1 3「6 5110\^11 111

0111—3.8.

001(301 3.8 7116 ! ) " 皿 ? 1:0066111:6

1)7 匚1&33: [ ; 糧 4 7 01833: (;糧40

0^ 014025

1 14

0.002226 0.002188 0,0021丄8

6 6

01833: ^ ^ 3 5 ^7^13 5

3 10

0.002034

^ 1 3 5 3 : 0^30 。30

6

0.001694

0.001715

8 2 《1.1皿2 112 〈1-111^2

0 1 1 1 I I I 3 . 9 5 1 1 0 ^ 5 1116 !'05111(3 0 「 1116 ! " 0 1 ^ 3 ^ 5610011011 ^1~00655 ( ? ^ ^ !). 0\\\^\\\ 3.9

111& 0^11^: ?10(^6(11:1:6 ^01: 丁 & 0 化 " 叩 51110

0^

111:01: 116011

? V&!'46 01&33: 07*^47 6 3

0,014025 0.002005 0,001777

127.141908 3.033106 5,385909

0, 0058 0-0011

11.284807 1.613065 1.430059

\\^11611 ^011 00111(331-6 01!1|31115 3 . 8 出 1 爿 3 . 9 , 乂011 0311 5 6 6 11131 1116 01355 ^31~135105 〔 V 八 & 3 5 11161 II化1^1111111111111 11-8(111X1-6 0「1161^011 !11 1116 8111 111 1116 8 6 0 0 1 1 ^ 5【6|3, 0 1 1 ^ 1116 ^[-011

0.088758 0.088637 0*086506

0\^^\1140 211(1

5 1 6 ^ III 1116;「 0「;^1131 3 1 1 ^ ^:01\^6

^3^131?163 3 1 1 1 ^ ^ 6 丄 0 1 5 ^ 1 3 ^ 3 . 1 2 8 1 1 0 ^ 5 1(16

^31*13&1*18 \^1111^0\^ 0 「 0 化 1 1 0 X 1 110(^6, ^ 1 ^ 1 15 II化1^6^1^6881011 , 1 0 ^

0:5^13^ 3.12

1156 0^31111 6,6^35 0 (:職40 哪

「化'1^ ^0 ^0

,叩"1 1叩& ^卩化

1^6 \'& 1

',

18^6!

^0

&0咖31 8 1 ,

21

^1)1


3.2.4.2

0(11-30)1131-6 0^1161-100

115111^ 1110 531116 ^ 3 1 3 561, I 5610016(1 1116 0111-3(11131~6 01~1161^1011? 5110\^11 III 0厂1116 7 3 1 - ^ 6 1

1^10(161

1 5 1 - 0 ( 3 6 ^ 10

0111~8^1131~^

0:3|3137 3 . 1 3 沂 1-561 ―0,16(1 03(3 ^ 〕 ^ " 0318

國,.

、力「13匕165

01383匕。1/61

100 50

731-9611^01^61 1-11016 ^6^601611 73['!31)163 6(601 0111156^ 乂3【131:155

丫65 丫55

| 卜"01111)6!" 0(81115

50

^3X1^10111

6

3 3 3 ""!!!匕"

~~ 1^1111101001 0^11-501131-8 同 6 - 3 0 賺 6 0011005 1-^3X11711^111^31-131)16他他6!"

3 . 8 4 ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ 3000

!-删1111^1 ^ - 5 ^ 3 「 6

0.0020

1-510^ ^ 5 [ ! 1131-6

5.0^-4

;1156 ; 1 ^ ! 6^3)131^165

1^0 丫6 5

; 1 1^56 1018 「31:110115 \

8?08

0 卜83

:1-1111 0011011

5-1^110 1。

11/23/06 6:19 ^ 3205353^22~47112~&160-70

1-1-951 813(113

00111 !]1@16

! ^ 8 8 ( 1 8 1 0 ^110

1-711116 0(13511^1111

111^23/06 6 : 5 3 ^

卜 ^ 1 1 0111-311011

;0

| :'11016 0^ 0^83(100

011)111.16.30 3 6 ^

'-01\^ 1-1051

0;5口 1 3 ^ 3 . 1 4 5 1 1 0 ^ 5 1116 :1~66 001151『11^63 !))'

!!!^^^^.

3 ^ 1 3 , 1 ^ 5611111^ 1110


98

?^1011^6

1^0^0(111^

6,7^^6斷。':?)^01^1 80111(10/15/(?/' 811^1^3 ^^1100110)1^

848

0 1 印 1 3 乂 3.14 了^',^^[!^!!!! 顯^]#101^1

0

0 :

0 0漏:!0 0*1/!'3 1

1

^ | ―匸―

^ 0 0 ^ ^3X8

^^135 5

10叭0^018

^,.

0^1, 2.3

1,2

0^*1511 3.1

4.7

0730 1 . ^ ! ^ , 0*^78

^ 0 0 ^ ^101)1

1116 11311165 0 「 1116 ^31~131?165 5 6 1 6 0 ( 6 ^ 3口口63「卜6|0\乂 1卜6 111161^116^1316 ^ 11011-101'!11111&!) 1 1 0 ^ 6 5 0『1116 1^66 111 0!51313^ 3.14.

015口!3乂 3 . 1 4 八 5 | 1 0 化 5 1116 ^31~13^16 11111501131106.

0:3|313^ 3.14八

| 「116

[ ! ^VI洲 别 ^ 。 "

|5~

[!

論 訓 職 川 屮 " " , 1 懼 丽 " 勵 ^ 疆 ' 爾 " 『 續 : 國 刚 8 8 0「一: 1 3 1 2 ―"1

^30 1

∼ 肅 8

1 1 1

1 0.453659 0.342075 0.311563 0.262701 0.262662 0.224586 0.197157

^ 1 6 0 乂011 001111)3!"6 111(3 乂3!"1^)165 ^1^011 111 0 1 5 ( 3 1 3 ^ 3 . 1 4 ^ ^ ( ( ! ) 111056 III 0111|31|1:5 3.8 011(1 3 . 9 , 丫011 0211 5 6 0 11131 0 \ ^ ^ 1 1 4 7 6 1 1 1 6 ^ 6 5 3 5 1116 111051 1111^0113111 乂3「131?16 6 丫 1)0111 1116 1 1 - 3 ^ 113^0 311^ 0111-3(11131*6 01'!161"!八5 ^011 111^111 「61:311,

^ 3 ^ 1 3 5 1 6 15 311 ^ ( ^ ( ^ ( ( ^ 0『11116「1161:

7 1 1 6 5 6 VIII'!II 1^165 10 0 1 5 ( 3 1 3 ^ 3 . 1 4 八 3 「 6 6 X ^ 0 1 1 6 ^ 10 11化116X1 1,03已1|1 11115 6X3111 [^!^, 1116 116X1 0 0 ( ^ 6 !5 1116 14.0^1-0881011 110^6^ 8 ^ 0(30111112 115 V;! 1*1111)108 \^1111^0\^5 乂011 XVIII 5 6 6 1116 531116 ^ ^ ! ' ! ^ ! ) ! ^ ^ XVIII! 1)10 1^010 0^1111111^ 丁 116 V;! 1-15*1)168 \^111^0\^ 0『1116 ^ 6 ^ ^ 6 8 8 1 0 1 1 110(16 15 5|10\^11 111 0 1 5 ^ 1 ^

3.15.


0 — 1 3 乂 3.15

^ 6 0730 0^0^ 0^20 01^34 07^40

| 1)36

战 1叩"1 1叩-"' 1|1睡

^0 110~ 1140 ^0 ^0

:。"31111 06131111 06^1111

06^31111 ∼

丫65

10^61

0116

1^6

1^)01

厂"II!"

1:細11

1叩」1 1,1

^01-1111131

!11(101

8111317

(:舊34 (:漏35 ^ ^ 4 0

丁3「。61

11

51^111 乂3「!313|65

5 0 1 0 0 1 ^ 5 ^111011 ^61^6 111056 11131 1)1-0(11106(1 11)6 1)051 5(51115 111 1116 1^^610(31116111

0「 1116 IV&&. 丁 116 36160(6(1 乂3「;31?165 31*6 ^ 6 1 1 11)6 1^016 。"叩!!^ III 1116 ^6^1*688100 110(16, 1186(1 116\^

3.3

丁 ! ^ ! ^ 。 ! " ! ! ^ ! ! ) ! !

3.3.1

0 , ^ 3 1 - 1 3 1 ) 1 6 3

71-3115101*111 ^ 3 1 * 1 3 1 2 1 6 8

丁1化丁『305『0「01

&1)163 00(16 0&II

^10^6

1113^6 3

乂 3 ! " 1 1 1 6 5 6 \^61~6 ^ ^ ^ ^ ! - ! ^ 111 86011011 2.6.1.1

0 『 1 「3115「01*1113110115 0 ^ 111(61^31-30&16(1 ^ 1)1 30^^1011,

31^6 1\^0 1^(365 0 『

11-3115101*1113(10115『0「【116 0316^01^1031 ( ^ 化 ) ^ 3 1 ^ 1 6 5 :

鲁 0 。 叩 , 0 1 1 1 1 1 0 1 丫 1113&310「5 丁「3115『0「出31101, 「0『1116

1『^011 \ ^ 3 0 ! 1 0 I!"3115^0!"111 ^31*1&1)168 ^\011。 1)3(3 1 1 0 ^ 6

86011011

\^311&1)10 5 6 | 6 0 1 1 0 1 ^ 乂011 0311 5 6 1

2.6.12^

1 1 1 6 卩 & 也 3 5 III 1)111

1 ) 3 ( 3 ?31*11(1011 110(^6 ― 丁「303!"01*01 ^^「131)108 110(^0 ― ^^!「131)16 8^160(1011 1 1 0 ^ ^

3 5 5)10^11 III ^ 6 叩 ^ ! " 卩 3 1 卜 I I I ! ! ) — — 3 . 1 6 . 1『乂011 (!。 0 0 1 \ ^ 3 0 1 1 0 1*6^601 I。口 11(8 0 0 1)16 1)3515 0 「 11116&I1 1^61&110113111|3 3^011&, ^ 0 1 1 0 3 0 1156 11118 ^1*00653 1 0 03)3111 !'6 11011-111103!"

&110115111133 1 ) 6 ^ 6 6 1 1 出 6

!11(51115 311(1 1:116 1 3 ! ^ 6 匕 9 0 ^ 6 ^ 6 ! ' , ∼ 1 出 1 3 ^ 6 6^1^ 5015 11115 ^ 0 0 2 5 5 1113^ 1)0001116 ^11110 1 6 出 0 1 1 、 XV11611 11161-6 31^6 3 1^!"^6 1111111136「 0『111^)1115 (巳经.,2000 0!" 11101*611 1X13^ 1)6 1110!'6 ^ ^ ! ^ ^ ! 1 0 0111111)1310 1 ^ 1 ^ 3 1 1 1 ^31-1315105 0『51, 31化111011 ^ 0 ^ 1 1 6 I「3115「0「11131;005 011 1111(30113111: ^31~135165 0 1 1 ~ , 丁卜6 131-02655 0 0 \ ^ 『 0 「 1|1;5 15 III 1)111 0 ^ ( 3 1 1 0 ^ 6 》 0 ^ ( 3 ?!11*11(1011 110(^6 ― \^31^131510 86100(1011 1 1 0 ^ 6 ^ X!'308101-01

3.16. I ^111 6X3111106

1*1115108

丁 1^5 36(^1101100 15 1 ^ 6 5 0 1 1 1 6 ^ ^ 1110 10\^01~ 口 3 1 卜 0 ^ 0 1 5 ^ ) 1 3 ^

1116 「6511115 0^1)0111 0 「 111656 50(^11611005

XVIII!

31(0111311^6 I!'3115^0111131:10115 115111^

3 5111^11 ^ 3 1 3 561 ^ 0 ) 5 ^ 1 3 ^ 3 . 1 6 ^ 0 6 5 1101 1110111(^6 1116 11111)11(0 11036, 1^111 III 3 0 乂 0 ^ 6 1 1 151-0)601 II 1111^111 1)6 0 6 0 6 5 5 3 1 7

1 0

" " " 汉 1 1 1 1 5 110(^6 (^『。!^ & ^ ! ^ ! ! ! ^ 1「3115『0「1113110115 0!" 36|601111@ " 「 ^ ^ .


^!0^11^

100

^1(11 848

&116?^156

1、!1訓':^0011001 8011(110)15/01-

81(51)1635柳11(^0纖

0:3|3137 3 . 1 6

翻3|3|6 3616011011 ^31

丁【303【01711 731-13^165 (^)

?^691-6551011②

丁 0 1^611101151!"316 1卜0 I「3II5『011113【10115, I \\?1\0: 561601:6^ 3 501311 1111】11匕6「 0『1(1 口 11(5 311(1 1116 1^1131^ ^68?

5

^111011 「6口『656018 1116 ^5|301150 1 0 3 5 0 1 1 0 1 ^ 1 0 1 1 \ ^ 3 11^ ^0111611031 I ? 殳 . ^ 6 1 * 6 01*6

111656 1叩1115,∼1"(;!1 3「6 5110、乂11 !11 ! ) ̶ "

3.1 7 :

11111111161- 0卩011̶1)116 111^1111'!65 1 ^ 1116 01151011161- |3611 111011廿1 11111111)6!' 0?八丁!VI ^6(305113

乂 八 ③ 乂 八 ! ^

11)6

乜1「66 111011(115

0『八丁!VI \^|1|1^1^3\^315 ^111*1112 1116 ^ 3 5 1 1111^6 1110111115

1111111

0^ 3 5 5 6 1 1^^)65

乂 八 鼠

311 !11(110310!' 0 『 & 1 1 1 1

乂 八 眺

11(6

乂 八 動

1:0131

5^11601: 0^1)31114 1310^11015 0^116(1 ^

01-6^11 500!'6

0)5^)137 3 . 1 7 ;

― ^ ~ ― !

【化『3

0(0?

1.8761

卿 1 ~ 1(1(6^1 ^ 2 0 111^111 14011)11181 柳 3 4 一 1叩!^ ^011)1(131 ^ 4 0 1^0111)1131 ^ 5 0 ,0^1―∼―― ^01)111131 ^ 6 0 ^011111131 (打)【111【131 糊80 1叩111 丁 【&50 8111317

^1

|

^0 110 吡

^0 。0 ^0 ^0 ^

140 80 80 ^0

0

^0 ^0 附

±1

1|化01! 5101116!"


010/711'/- 3:厂"〃'"/,/。 5^1601 !017 0)1^

3.3.2

丁!'∼"/(",而!!。"

0^ \^01~10^!^

丁 ! ^ ! ^ 。 ! 1 ! ! ^ " 。 ! ! 136,0『6 ^31-131)16 56画6。1:。11

1|1 1116

^1111 0 ^ 0 1 5 ^ ) 1 3 ^ 3 . 1 8 11化^31-13(3165 ;1「6 I「3115『0「11^3 115111^ 1116 X 1*308101-111 \^31^1111)108

1101^0 311(1 111011 ^ 3 5 5 6 ^ 1 0 1116 ^ ! ^ & 比 8 0 1 0 0 1 1 0 1 1

||1 11)6 71-308(01-01 VII 1*1111)108 110(^6, 1 5 0 1 0 0 ^ 1 ^ 1110 1^3X111111111

1^01"!!"II

1「3115『0口11311011 3 5 1110

^6(31111 171611101^ 「01- 3!1 00111111110115 乂 3 「 ; 3 | 3 ^ 丁 0 1113^6 11115 5610011011. I 501 11)6 111161^31 1 叩 1 1 ^ 1?「0「611^ 1 0 1^13X10111111 1

^

0

"

I

1X3II5『0「11131 !011,

110(16

1 1 2 1 - 3 1 6 5

^1^0?"6111

卩0\\^「 1「305厂0「1,,31;01,5 0 「 1(16 乂 3 ! " ! 3 1 1 ^ 口1(^5 11化(!"3115『0「0131;011 11131 1「3115『0「1115 1116 111^1)1 11110 3 0 6 ^ ^31*13^16 \^II086 3!31!"11)011011 15 0 1 0 5 6 5 1 1 0 110!"11131 1)611-5113^^ ^151 ! - ! ! ) ^ ! ^ ! ! . ( 化 ^ 。 。 2.6.7, I ^ ^ 6 5 1116

0「卜0、乂 II 00111^31~65 1116 ^!5:1~113111!0115 0『311 1「31,5尸0「11131|0115 ^ 1 ( 1 1 出 6 II01-11181

爿!511";1)1111011 311(1 56100(5 11化0116 0 1 0 5 6 5 1 1 0 !^.)卩0「 0 1 3 5 5 11\1115^01111311011 ^ ) / 5611111^ 1116 0 1 3 8 5 1(1)111(8

I 5 6 1 6 0 ( 6 ^ 1)16 1)11111111^ 111^11^(0[-^

!'0^611^ 10 011111111^】11(1;""0『1 7 1 1 6 5 6 56^11155 3「0

5110^11 111 0 1 5 卩 ― 3 . 1 8^

0;3|313^ 3.18

III 111(5 0)313 561 1|161^6 31~6 011 ! ) 1 1\^0 1 1 1 1 6 ^ 3 1 - 5 0 3 1 0 ^ \力「136165一 丁 116〉, 3 1 2 1「 3 11110161-10 \^\3^\6

311^

011^0500^^

1138 0 0 1 ? 3 「 6 ∼ ^ 1 1 6 3 , II化11 6111617)1*156 ^!!06!" 36(3 1(5 0163511^6111601 1^61 1 0

^ 0 0 1 1 0 3 1 I丁 1116 !\16^(1:11:1八^1^150「 01)1100 15 561 1 0 八 3 ^ 0 0 6 ^ XV11611 01*631 !115出6 ^ 3 1 3 80111*06 (^^ 0;3口13乂 2 . 1 4 八 ) . 乂 八 1 0 ^ , 乂 八 尺 2 0 , 乂 八 尺 5 0 , 3 1 1 ^ 乂 八 尺 8 0 3 化 0 「 山 1 1 3 1 , 3 1 1 ^ 3「6 501 1 0 ^011111131 !?)' 511121^1-156

丁 卜 「 0 「

11化^「136165

501 1 0 !11161^,3卜5(^16(1, 1110 111X16 5631^1化5

311(1 0 1 1 ^ 5 1116 1^3x111111111 1^01*0131, 3 5 1 5^601(10(1 23「116「, 311(1『0『1116 「651 0『1116 乂3「1313|65,"'卜&卜 31-6 1 1 0 0 1 1 1 1 1 1 3 1 - 5 0 3 ! ^ , II 01^63165 3 (!^!!!!!!)' ^31*131)16『0「 63011 ! ^ ^ ! . 丁卜6 1 「3115「017116(1 ^ 3 ^ 1 0 5 1 6 5 3!"6 111611 ^355601 1 0 1116 116X1 110(^0 III 1116 |31*00655 (!。∼.1|1 11115 6X3111 口!6, 1116 ^31*131)16 861601100 110(^6 ^ 0 1 1 0 ^ 5 1116 X1-3118101*01

1*131)108 1 1 0 ^ ^

^ 0 1 1 0^)611 1)10 ^31*131)108 \^111^0\^ 0^1116 ^ ^ ! " ! ^ ! ) ^

80^011011 110(16, ^ 0 1 1 0311 5 6 6 1116 ^3^!411)165 | 3 3 5 5 0 ^ 1 0 II ^ 1110 X 1*3118(01111 V I I 1*131)108 ^31*131)165 31*6 5 ! 1 0 環 I I I 0 1 5 | 3 | 3 ^ 5

3.19^

311^

3.198.

711050

101


013(313乂 3.19众

1)36 06^81111 ^68

「65[! 丁1 一 卿 2 0 1 !'1^/^20 2 71^^203 11^^204 ― II , 2 0 6

1.0761

8。|8

0

1(1161731

1)6(31111 06(31111

731-961 , 1叩口 1 1叩"1 1^1)1 111(3111

06131)11

111(1111罾―― !811131^ 81113^

06(31111

111(3111 1,1

8111317

8103^

卜1

61(13^ II卿208 了,)八細10 71^^3^111 了 1—^63^112 丁匕霍31^13 丁 1—窗3门14 !"1^^3^15 11—^31^18 !"1一狄賺17

6103^ 81(1317

爾 效

1(1)101 ;

^

^ 61113IV―—一'

06(31111 ―….――"

06(311(1

:313∼ .1叩1)1 ,叩"1 ^ ^ ^ 81(13^ 111 !] 111 81113^

。6^3"!1

1(1)1111 1,1

7 1 训 閥 2

卜1

8111917 曰1。3∼

― 爐

了―1一卿3∼4 多̶31^5 ~ 7 1 ^ ^ 3^)6 71 " 閥 了 7 1 讲 謂 8 71^^31^19

06(31111 06(3"!1

10(3111 1,1 111 !]^1

06!"31111

^(!^!

^ 7 「

8111317 8111317 81113^ :81啊

,,

\

,1 0^

0:3|3|37 3.196 71^^501 丁匕卿502 06^3"!^ 丁I额504 VI職505 71^^506 71^^507 1-1^/^508

^311(1

1门^1 ^'―.」!| 1111111^ 11^111 11^111 1,1 1,1 1,1

0&131111 !'1^^804

^ " 關

^ 4 0 ^ 6 0

14 61啊 81啊 8|11317

~

81啊 81113。 81031^

111(1^1 申 1

71—^5802

71^^806 71^/^807

1^

8103^

06(30(1 一&的!!"

1^

8103^ 奉 1 1^111 111^111 1叩111

1,1

8111317 『

1^16^31

106(31!" 06^311)1 0 # 誠

1^011111131

0

^6)80181^

11

^| 一。『6一 上

|| 0 4

|

080061


0/10声"3: 701-101)16 86160(10^1 0^

?!'001 0!5口 1&乂 3^ 19八,II 0 3 0 1)6 8660

丁"。"圩。"賺I!。" 0/^0^101)168

& 10@ 1『305『0011811011 1113016 1卜6 0 1 1 5 0 8 0 0 1 1 5

^\&1)1&

010565110 1116 00111131 (!&^!!)"^。!!. 1^0 11*3113^01111311011 00111(1 1)6『000(1『0『\^八^11^1; 116006, 11 13 1)3356(1 10 1110 116X1 00(^0 3 5 18. 1『3 1「305『0『0131100 15 1113(16『0『3 乂3『13|316,出60 ^

^ ^ 1 1 1 1 001乂 1卜6

-

II &031011116(1 ^31-13616 13 ^3356(1 10 1116 06X1 00(16, 0011116 01*1^11131 \ ^ \ ^ \ & . 1『乂011 561 1卜6 111(16 151^0|3611^ 10 ^ 0 &1\6 1*16 1^6|60|; |3!^0|5611^ 10 1^0,出60 1)0(11 1116 01*1^11131 3 0 3 加 ! 1 5 『 0 0 1 1 6 0 1 窗 ^ 1 6 5 1)6 | ^ 5 5 6 3 10 1116 116X1 110(16狄!出1116 ^ 0 1 6 56110 11111111 丁 116 ^1101111^ ^3.119^1^5 01I&6『0『1116

01&55 乂3『1&616 16乂引5 0311 &150 1?6 36611 111 0|3|3|3^3 3 , 19八31101

3 . 1 9 8 . ?0『6X3011)16, 乂 八 1 1 2 0 1133 613111: 0316201-163. 丁 116乂 31^6 0 , 1 ,2,3,4,5,6, \^13^\65

(^&^

8 , 1^6006, 6\^\\1

V3I^13^)I65) 31*6 01*6316(1 ^ 丁 116乂 31*6 丁1一\^八11201 1111*011311 7 1 ^ ^ 1 1 2 0 8 ^

丁116 乂3『1&1)16 乂 八 1 1 2 0 1135 50016 1111531113 ^ ^ ! ^ . 90\^6乂6「, 0 0 56口&1^16 0 3 1 6 ^ 0 1 ^ 13 0^^3.1&6卩0『11. 丁!^!^&化,00 311111111 乂 ^31*131)16 13 0^6^1661卜6 丁 1—乂八^201 (&1(63 1116

1111351113 ^811163^ 丁卜6 30111111 乂 ^31*131)16

0 ^ 1 1『乂八1^20 18 0 , 31\6 0 011161^13^ 丁 1^\^八11202 0 ^ 6 5 出 6

0 ^ 1 ^ 乂 八 1 1 2 0 :5 1,311(10 011161^15^ 丁 1 7 八 1 1 2 0 3 01(63 1116 乂 ^ 此 0 ^ 1 化 ^ 八 1 1 2 0 13 2 , 30(1 0 011161^136, 610. 丁 116

&6165 36160(6(1 6乂 1(16

&1)16 86160(1011 110(16 31*6

^ 011中111 3.10.1^016 11131 0 0 ^

1^0 ^31*1&5163 ^61*6 5616(^601, 1110111(11113 0 0 0 0『1116 0 6 ^ 31101111 乂 ^31-131)165 01*6316(1 6 乂 出 6 11^05&1111&1100. 丁卜6

8 6 1 6 0 ^ 1 0 0 110016 01-63(6(1 1)16

出 6 1 1 8 6 八 0 ^ 1 6 ^^『131)168 ^1-0)3011^ 0『1)16 \^\^\^

& 6 1 6 八 0 乂 1 6 \^八1111^ 1)6031136 I 361

8 6 1 6 0 ( 1 0 0 110(16 10

0111^)1113.10 1116 1)111112 ? 100^(1111:6 ^11^0X3 011036X1 11^

7 & 1 0 6 1 : 6 3口 3101 0 2

? 7&1116 2 14

&0716: 7 ^ 1 9

丫011 01&乂

0.007478 0.004542

33.667268 2.930341

:101: 1160X1

3(111&169 《0001 0.0002

6.016813 3.654798

0.089357 0.089088

0 1 1 0 口355 0101*6 111811 0116 ^ & 0『1^305^01*11131!00『0『63011 10161^31 10|5111: 10 1116

^ 8 1 * 1 0 5 1 6 86160(1011 110(16 01-1116 ^631*6881011 0 0 ^ 6 1)1 01^6『10 361601: 1^16 1?6116『0『1116 1^0, ?0『 6X3111 口 16, 乂011 0 1 3 ^ ^ 3 0 1 10 |3335 1^3X111111111 ^01*11131 311(1 0^11111311『3115『011112^0115^ 丁 1115 0 3 0 1)6 30116 ^

3^1^1113 1\^0 1113011063 0『1卜6 丁 1^08【0『111 ^ 8 1 * 1 & 5 1 6 8 1\06&. 1|1 1卜6『11*81 1113(31106

( 坨 口 ! ^ 加 ! " 6 ^ 71*308 111

3.20〉 0『1116 71*308(01*111 \^31^135168 00(16, I 11*3113^61111 &II 1卜6

1)1161^81 11^1118 1131113 1^6 1^3X111111111 ^01*11181 (!1&115『0『111011 ^

361(1113 1116 111161^31 1 0 1 ) 0 ( 8

^1*01361^ 10 1^13X111111111 ~0『11131 1|1 1\\6 560011(1 111301106 ( ^ ! ) ^ ^ ^ ! ! ^ 5 ^ 丁 1 ^ 0 8 2 III

丫 3.20〉,

I 1^305^01111 &11 111(61^31 111(311(3 1151113 1^16 0^(111131 11*3115(011113(1011 ^ 36(11113出6 111161*^31 III |111^8 |31^0|3611^ 10 0|^!11131 丁11611, 1131113 1卜6 8 ^ 8 0 0 ^ 6 110X16, I 016!"^ 1116

& 5613 01*63(6(1 1?乂 1116

丁!^!^&!1!!! \^31^186168 110(163 3 0 1(131 I 0311 口355 1)0(11 1^)365 0 ^ 11*303&1111311011 10 ^ 6 ^6^1^6881011 00(^6『0『厂1031 36160(1011 0 ^ 1116 ^31*131)165 001151(161*1113 ^ 5 0 出 ^ ? 6 5 0『11*30 3^001131100 ^ 0 1 5 ^ 1 3 ^ 3.20 3110^3 1116 口『00655 110\^『0『001111)1111113 1(16 1 ^ 0 1^(565 0 ^ 1|^115『001131;0115,

103


104

^1-^10(1^

^0(1&11)1^

1177/7 5^5

&710^!1^1^邊16!': ?)~001\00\ 50/11/10175/。!' 81151)16^

/―!^!化,化

015^13^ 3.20

81 !^^8^^1351

丁 116 1^1^X111111111 1^01*11131 112115『0「11131:;011 15 ^0112 1^ 1116 X 1^3II810111 ^31*131)168 110(16 ^ ^ ^ ? , 31101 【I化011110131 11-3115^01-111311011 15 ^ 0 1 1 6 ^

71*3II8101*01

15 115001 1 0 001111)1110廿,6 ^ 3 1 3 5 6 ! 5 ^ 1 1 6 1 ^ 1 6 ^ ^

1*11115165 110(16 X 1*3082^ 丁!16 8 X 8 0 0 ( 1 6 00(16

^ 1 6 5 6 1 ^ 0 110(^65, 311^ 1116 00111^1112^ 6 ^

561 15

|3355201 1 0 1116 1^021*6881011

015|3!3^ 3.21

5 1 1 0 ^ 5 1 1 化 3 八 5 00(16 1156(1 10 001115)116 1116 ^ 3 1 3 56(5 01*63(6^ 5 ^ 1110 丁「305『0「111

1-1111)165 0 0 ( ^ 6 5 X 1*308 ^111^ 1「3!182^

0:5(313乂 3 . 2 1

! …

加 3

55^2^?0 ^ 1 ^ 1 * 1 ;

―33*3!:乂?。&丁—^110;!!!:; 磁 " 4 化 」 1 6 . . 1 刚 3 — 7 ^ 1 1 1 ^ ^ ― ! ^ , ,11^132—^111^1^

!161:96化!^^]^.

11:3丁

8 ^ 3 00^6苎 力^

1^016 11131 1116 1^3(3 8615 3「6 1 1 1 0 ^ 6 ^ XV 11)10111 3 8 ^

51316111131,I, 1)60&1150乜16「6 15 2 0116-10-0116

001-1-6815011(161106 1 ) 6 1 ^ 0 6 1 1 出 6 1*0001*^8 0卩1116 ^ 3 1 3 5^15 1 1 1 6 ^ ( ^ . 丁!16 01&01*0 ^31*131)105 0011131111113也 6 11311105 0『1116 6 X ^ 0 1 1 ^ 3 ( 3 5615 &1*6 01^316(1 6乂 1116 8 八 8 0 0 ^ 6 ! ! ^ ^ . 1^010 11131 11化1^0(16 1 0 口『0口61!》' 0『1116 111-51 1115131106 0 ? 1116 X 1*3 08*01*111

1)168 1 1 0 ^ 6 15 1~1~3118, 3 1 1 ^ 11131 0『1116 3 6 0 0 1 1 ^ 1115131106 15


01率(^!' 3: \^11106!6 8^1^0110)1 011(1 丁,,和.瞧!!011 0^ ^01101)1^;

XII6

& 1 6 8 ^111(10^^ 0卜1116 1^6^1*6881011 110^6 15 5)10*11 111 0 1 5 ^ 1 3 ^ 3.21.

0:3卩13乂 3 . 2 2 她1訓,^^!^1 ∼' ^ 0 | 11^0 100」『5(^0【5 0^2111! ;^"」 0^3111^ ^ ! ^ " ―

| 3013011 ^0 ^0

| 17010 1叩"I 1叩"1

^0

1叩"1

^0 |40

1,【 ∼《「丄'3

0"城 ^60 ^^70 ^60

\ 16^ 1门^31

\ 1^30 ^1

0 ^

1?

~0巾1。31 ^ 6 ^10^31 1^ ^ 『 幽^ '…,1^"… ^001[[!31 ―,4舰^1―""!5一

| 丁^5(0—50: 0\66\\ 丁【303(0^3(1

― : 、

∼0《

0

06(3^11 140 1)61續 ―》? ^10

1啊【 !^) 10^31 丁 1061

1\

丁 1化11-3115^01"!113110115 1^0116

1)0111【I化II"!1118『0「111 \^31^131)105 0 0 ( ^ 6 5 (^!!^^'!!

0 0 ^ 。^11&1516 1 0 1116 ^ 6 ^ 1 * 6 8 8 1 0 0

0^^05001^5 01-63(6(1

?!'0111 018(31&)' 3 . 2 2 , 11 0311 1)6 52611

1135 1^/0 ! ^ 口 " 。 卩 ^ 化 & ! " ! ! ^ " 。 " . 丁 [ ! 6 0 「 9 0110 15

1)1 013^13^

3 . 2 0 〉 3「6

【!16 、 0 1 ^ 種

1^00^01150500^55∼"^

1「3115「0!"11!"1^ 1116 乂3「;361。 0「6^1化0「0 115100 1|化1^111X111111111 1^01*0131 11161110^ ^

^「^1)16 01^203001^5 01120500^5 1116 0^1111:11

XI'&08101*111 ^31*131)168 110(16 X I 1 1 8 . 丁 116 360011(1【「3115『0「111:1^011 0『1110

0?1^0^03001^^

0!~6^16^

5^

113115「0「111!1巧1116

15

1116 15

1^3II5^01*111311011 11161110^ 11131 15 口6「「0111化3 XVIII! 1116 X 1*3118(01*01 \^31^1111)168 110(^6 丁『111182, 1|1 31^^111011 5 1116 0「1^!1131 ",化!^'」乂3「131?16

01180800115

15 3 1 3 0 ^ 3 5 5 6 1 ^ 1 0 1116 ^ 6 2 1 * 0 8 8 1 0 1 1 1101^6 XV1111

115 1^0)6 3 5 5 1 3 1 1 6 ^ 1 0 101)111 511106 I 561 1116 1^1110 3 门 3 ^ ^ 0 0 1 !3「0|3611|25 1 0 1^0 !11 60111 1116 71-3115(01*111 \^31^!315108 110(^65 X 1*3118出13 X 1*11 II82^

\\^6 0311 & 1 3 0 5 6 6 1化「6

1)16

'!&1)10 \ ^ 八 6 1 ∼ 卜 3 5 01、1 乂 0116 1(111(1 0 「 I!"3115^01'!11311011. 31\6 11131 II 15

3 0 0 ^ 1 1 1 ) 3 1 11-3115^0011311011 ^ 丁 116 1^3X10111111 1^01*11131 1116111001 3 ^ 1101 口!"。""(;6 3 0 ^ !^3115!"01111311011 0 『 乂 八 & ! ] ^ , 1)603056 !!16 0^1@11131 V&!'131)16 13 0!056「 10 110「1113! 111311 311 乂 (^)"^!" 1「31150111136011 0卩1116 01 !^1113^

0 啤 1 1 1 3.1 1 5!10、乂5 1116 ^ 「 1 3 6 1 6 5 5616016(1 1 ^ 1)16 " ( ; ^ 6 5 5 ; 0 1 1

\^31~131516 5616011011 ^

1116

0口1|3口1 3.11

1^)6

3 ^ 0 3 1 7 3 1 3 。1 ^ 6 ( ^ 3

靈 8 0

1^016 11131

\^八1180

0?

0111-5(11131:6

6

20.7430 15.1017 13.2491

15 3 110111111&1 ^ ( 化 ^ .

? 1; ^

^

《0001 0.0005 0.0056

丁 1 化 0 1 3 5 8 111^)11(5 口 「 , 化 「 0 ! " 11115 0 3 5 6 ^ 3 5 5 6 1 1 0 1 ^ 0 0 巳

9 6 0 0 6 , 1116 V 1 ) 1 0 ^ 3 5 001: I: !"3115?01'!!!^^.


106

^6(11011^6 140(1&11^

3.3.3

& 4 5 ^11(61-^1-156 ^11761': ^00(1001

71*3118101-111311011 3 # | 6 『 ^ 3 1 * 1 3 1 ) 1 6

80114(10115/(^ 8143117683 ^^1100(10118

86160^1011

I犷乂011卜3乂6 & \ 1 ^ 6 口"!!!!^!" 0『^81*131)165 III ^ 0 1 1 『 6 ^ 8 6 ^ ^011 01&)/ ^ 3 0 1 1 0 ( ! ^ 6)11111113(6 111036 11131113^6 6乂1『60161^ 1 0 ^

001*1*613(1011 \^!1|1 1116 ^ ! " ^ . 丁 0 30111^6 11115, I 5611116 乂811165

0『1116 1\1!0111111111 1^-8(11181*6 311(1 ^ ^ ] ) ^8(10&1*6 口『0口611!65 0 ^ 出 6 ^31*131)16 86160(100 0 0 ^ 6 10 0.005 311(1 0.0005,

^ ^ ^ ^ ^ ^ ! ^ .

丁 116

^31-131)163 36160(6(1 111 816^ 1 0『1116

86160^1011 31*6

5110\^11 1(1 0 啤 0 1 3.12, 0 却 I I I 3.12

1^-3(111(11:63 201: 丁&1:9化^8^181)16:"印

01&33: 观3议 ! 61: 0X111: 01883: 7 ^ 5 0 1 0110^: 7 ^ 5 0 01&33: 61: 011。:7 ^ 2 0 :01&33:7 ^ 8 0 I !)"。!): 職 8 0 ! 層 1 6 2 ^11111

0?

11-5(11161:6

17 5 8 3 8 4 6

0,010908

2

14 1 3 1

乂":

&0716: 016(130016 781:: 0116(13001:6

0,010724

0 0 0 0 0 0

009589 009517 009424 009410 008450 008392 0,006479 0,003508 0^ 000289 13.000162

1^2 ^ 11113^2 82 ^ 11111112 ^2 ^ 11111^2 〃」

:'-"、:-∼一.';~ ^ ― 一

0111(3111: 3.13 5110\^3出6 ^31-131)165 56160(0(1 III 310|5 1. 0\1\^\\\ 3.13 7116 011111^ ? 1:006|1111:6 52260^3 01103611 ^01: 丁&【06仁:163。 311111 02

| 61: 011)1:7^311 | &0716: | 01: 01111: 啣20 ;0^0\1^:懂 8 0

0?

只-3^11&1:6

5 14 3 4 2

0.010724 0.003912 0.001891 0.000673 0.000567

? ^1116 19.369933 2.529776 5.715847 1.525318 2.574017

11-78111X6 《0001 0.0013 0.0007 0.1918 0.0763

8.628783 3.147905 1.521686 0.541304 0.456572

" 一 ̶ 一

015(313^5 3.23八311(1 3.238 5卜0\^ 1)16 乂3『;36163 ^3383601 10 1116 116^1*6851011 110016.

& 1:1:01: 06811 5(111&]:6 0.089095 0.088882 0.088741 0.088720 0.088689


015^13^

3.23八

「 " 「 亡

|

1^0

丁1 一 ^ 魏 7 1 - ^ / 1 &墨洲1 删 II ^ 0 7 1 6 ^ ^ 1 ^12 ^413 ",』 1 1 ^ 0 ^ 1 6 墨 ^ 1 ^2 ^3 ^4 丁 1 ^ 0 7 1 6 一 糊 ^5 71^0^10墨卿 ^6 1八。VI6一湖 ^7 71一八0716墨\^ 71 ^0^16

015(3137

|

\ ^

1

1

1 ^

0 ^

|

101)01

^0^16柳川:1 ^0^16 10 ^0 ^0 ^0 [ ^ 城 [ ^ 滅

―"。^ 110 !40

1叩!^ 1叩!]1 1,1

II

1^1 1^01 1叩111

^

1叩01

110 ^0 ^0 ||0

【叩^ 『叩"1 1响1 1响1

^ 0 ^ 1 6 7 ^ 1 ^ :13 ^ 1 6 糊 川 " ^ 1 6 7 ^ 1 ^ :16

61^3^

^ 1 6 補 1 4 : 3 01, 曰1门3∼

^0716糊1化6

6^3^

^ 0 ^ 1 6 湖 1 " :9

3.238

11 I 1136 71 3湖201 亍1—0 一^1202 71^0^^203 71^0^^204

|

03碑 [ ^ 城

0'化'

0赚20:0 0柳20:1 0卿20:2

【叩"1 111 口"1 1叩《

11一 0 ^ ^ 3 4 ,

061城 11 0 ^ ^ 3 1 ^ 6 [ ^ 滅 ,顯。一湖502

110 110 ^0 ^0 ^0 ^0 ^0 140 110

V ^^!^!^-^ 0棚20:4 0湖3化0 0 ^^3^:1 0 ^^3^:2 0 ^^3^:3 ^0柳34:4

,,1 1^1^

81031^

1晴1 1叩"1 1。卩"1 1叩"1 1^1^111

6103^

1叩11|

61卩3吖

61(13^

||| ― ^1

0 ^^50:0 0柳50:1 0柳50:2 0 ^^?50:3

061311^1 71—0墨\^?302 冊03

― 42 00 ^ 稱50 卿 5 0 ^ 7 0 卿80 口60"。& ^660^^^^^

061^

0(^31111

[ ^ 城 03城 丫65

^0 ^0 ^0 ^0 ^0 "!40 ^0

1 口16^1 爿0^门31

。5^0巳 犯乩∼叩0∼

^ [ 一 ^0

110^

0輕80:, 0卿80:2

、 1^111 ^ 6 ^ 1 6 [! 86拎""

∼0巾1门31 "(!1X11031 了 3/001

^

1

!7!"。'!! 0 1 5 ^ 1 3 ^ 5 3 . 2 3 八 3 1 1 ^ 3 . 2 3 8 ^ 0 1 1 0311 5 6 6 11131 1卜6 ^31-1:11)16 801001)011 1101^6 0 ^ 1 1 1 ^ 1)0111 1116

^ 3 1 - 1 3 1 ) 1 6 8 0 0 0 1 6 0 ^ 3 1 0 ^ 1116 ^11111111^ "「卜化!65 ( " " ^ " ― " 化 ! ! ) 『 0 ! " 八 0 \ ^ 1 6 3.1)6 ^01\^6

01355

^ 「 ^ ! " .

63^11

0『111056


^11(1x^2

1^0^01(11^

3.3.4

;177/1 5 / 1 5 1^1110-^1-1X0 \41^:

?1-00(1001 8011(11011!; ^

^^1100(101^

811X1/1^

33^1119 3 1 1 3 5 X 1 ) 0 1 ^ 1 1 9 仇 6 0 0 1 1 6 0 6 1 1 6 1 * 3 1 6 ^ 1 ^

1116 1『3115,0『111

^31*131)165 110116 丁 116 3 八 3 00(16 ^61101-310^ I,乂 1116 71*3118101*111

&1)168 110(16 0011 1)6 3 ^ 6 ( 1 & 1 1 ^ 2 X ^ 0 1 1 6 ( 1 ^ 5!111!?1乂

0|5011 1116 1^08111(8 \^111^0\^ 0111^ 5 6 ^ 0 1 V 比 、 V 5卜0職111 015卩1^

8 八 8 1^811118

卩 1 0 化 0 0 ( 1 6 ^0111 1)16 11101111 ^

35

3.23(1^

013^)13^ 3 . 2 3 0 ―-,―

^

――…一國…

「―――『―丁^―― ―

^

一一11

― ― :

― ―

― ―

^6511(15 ~ 了「80,『01711 ^31101)165 (力

1!]

11 1.0^

0"卬。1:

6 丁「305、 5001-06

「0!^敲,.

5 1 1 瞧 ^ 5^115^1:5》

1钆'"11161. 01 167613 :

"界"丁

15

0

20

1^11 丁

1^11 丁 50

1圖丁

咿11下 0 1 ) 1 ^丁 011则丁

01^1门3【 00111^1116^ 00^1^31116^

0 0 丁∼丁 00丁叫丁 0 0 丁∼丁 01)7^11 丁 01^「 口丫

0001|31|16

丁 1.40 VI 6—^/7^141 丁 1-40716-7/^11^10 了 1-4(^/16)^1411 丁 1》0乂16一\^11412 丁1一八0乂163^1门13 71^0^16^^11^114 71)0^/16一赠1门15 了 1—40716—7/^142 丁 1—皋0716—7/^11^3

2

2 2

0

':'

:'11

0^611

^ 0

II化1^10\\ 00(16 \^11|50化,5^1001 ^110

311^ 5 6 1 6 0 1 3 1131116卩。!' 1116 3 3 ^ 6 ( 5

^ ' : ^ ^

:;'::

II國八0^16—^7^1^15

0 0 0 0 0 0 0 0

4!'"

2:

:,

||01^「11下 權。11丁「11 丁

丁0 5 3 乂 6 1116

1

2

「「0111 1116 11161111

'…^ 一0

0 0

01101; 011

8^^6八1

丁 116 5 八 5 ? 1 。 ∼ 0 0 ( ^ 6 5 3 乂 6 3 111 111^ 6 X 3 0 1 ( 3 ^ 15 5卜0∼11 111

0 1 5 ^ ) 1 3 ^ 3 . 2 4 , II 15 1 1 3 ^ 3 1 1 3 6 1 ^ 3 5 3 (!!^ ?0「 6乂口011,


013(313^ 3 . 2 4

2 3

4

1611^:11 ^00^^0^0131: 5200;

5

一0 "墨?"13

6

011:015 ^00^?01:1181: 一011—^01:1181::

0 . ;

1 8 9 10

,,

,0'纖?

^01: 40716^\'&? III; 會8

3

,111111117 ^01: 1*^61 1;

11

191)61 丁 1—厶0716一乂;^111:1^ ^ ^ 1 6 ^ 5 : ^ ; 2 ':

12

一0!1一?01:0131; 2 ^ 0 ^ 1 6 ^ ^ 1 0 ;

13

11 4 0 7 1 6 ^ 4 1 ^ 1 1 1 巧 ~

1^1611 丁 1 ) 0 7 1 6 — ^ 1 1 1 1

14 15

1^ ^011^011)1161; ^ 1 1^1611 丁1—40716一乂^11】1 : 1;

16

6136 丁 1—710716—7^1X1

^ 0;

:二

17 18

101)61 丁 1—40^丄6一叮&!^!!之::'3(7/26 V^?,1^:2 ;

丄9

一0"—?01^81; : ^ 0 7 1 6 ^ ^ 1 1 1 ;

1

20

1^ ;10716一乂^1丄11 6 ^ ^ ^11611 丁1一;10VI 6—^4^1112^ ^

21

613&

22

12 ^011^^011118& : 2 ^1160 7 1 ^ 0 7 1 6 ^ 7 ^ 1 1 1 2 ^ 1;

23

6156 7 1 ^ 0 7 1 6 ^ ^ 1 1 1 2

24

,1)11111117伫0汇 161/21 3; 1 & 6 6 1 丁 1 — & 0 7 1 6 — ^ 1 0 \^16^^1^:

25

^ 0; 3 ; 1

26 27

11: 4 0 7 1 6 ^ ^ 1 1 ; 6(1 , 1^1611 丁 1—40^16—7/^1113

28

3.4

29

1^ ―011—^01:11161: ^ 3 & 6 0 丁 1 — 厶 0 ^ 1 6 ^ ^ 1 0 3 ^ 1;

30

6136 丁 1 — 皿 " — 乂 皿 0 3 ^ 0;|

31

,1)1.111^7 ^01; 1*^61 4;

31111111131^ #

71115 (:!化!^^^ 5|10、\,6^ 110\\? 1116 V I I 1*1111)10 80100(1011 110(^6 5 ^ 1 ^ 1 5 ^;31680「;1;31 311^1 001111111)0115 \^11~!^165 1 0 1)6 1156^ 3 5 !11^)1115 III 1116 1110^01111^ |31~00655 ^11611 1116 1 3 1 ^ 6 1 15 1^11317 0「 00111111110115.

^

丁 训 ( ; ! " ^ ! ^ 3!'6 3 ^ 1 3 1 ) 1 6 1^0!' 、 ^ 3 ^ 8016011011 ^ 1 1 6 0 1)16 1 3 1 ^ 6 1 !5 ^!11317: ^ 8 ^ 0 3 1 ~ 6 3 1 1 ^ 0111-301

^

丁 116 8.-5(111(11^0 0「!&「!011 0311 1?0 1156^ XV 1(11 1)1113^ 3 5

3 5 0011(111110115 13!"^!:^.匕III 1110

0 1 1 1 - 5 ^ 1 1 3 ^ 01'!161"!011 0311 1)0 1156^1 0111^ ^11)1 1)111317 1 3 ^ 0 1 5 ~ ^

丁 1,6

1*1511)10 8010011011 00^0忭3115「011115 111(31115 111(0 ^ 1 1 1 1 1 ^ ^ 3 ! " 0 3 1 1 ^ 八 0 乂 1 6

\ 1 11^136165^ ^111011 31^6 1.156^111 111 ( : 叩 " ' 「 ― 11011-111103!" 「61311。1151,1口5 1 ^ 6 ^ 6 6 1 1 I叩1115 3 0 ^ 13 「 3 ^ 5 , ^

丁I化V:"'111&16 80100(1011 110^0 3!50 51111 !')!11125 0310^01~10111 , ' ; 3 1 ^ 5 \ ^ 0011(1(55111^ II化

1 1 1 6 71-1*118101*111 V I I 1*131)165 110(^0 0^01~5 3 ^ 1 ^ 6 0厂「311^6 01101005 「0「 1巧1,5『0「1113110115, 1110111(1111^ 111111111516 1「3,,5『0「11131|0115 0!" 311

311^ 1卜6 0(511011 0 「 (;!^^""^ ^11111171^

、力「;油165 「0「 1116 0 3 ( 6 ^ 0 ! " ^ 5 0 「 031612.01-1031 " ! " ! ^ ^ .


110

? 1-0(110(1^6 1^0^1(11^

"'〃/,

^1116)^1-1X0斷16,':

1『11161-6 31-6 3

?)~0011001 80! 1/110115 /。尸811X1賺8

^^1100110)15

0「 1叩1115, ^011 0311 1150 II化V;! 1*1111)10 86160(1011 1 1 0 ^ 6 1 0 II

1113^6 3 ^!'61!|11|11317 ^3!"!31)16 5 6 1 6 ( ^ 0 1 1 , 311(1 111611 ^ 0 11*3115^01"!113110115 0^1116 561601:6(1 ^31~!^13|651 1『1116 11111111^「 0『!11^11(5 1 5 5 1 1 1 3 1 1

7

111011 ^ 0 1 1 0311 1「3115卩0口11 ^11 0『1116

1叩1115 !^';0「 1 0 ^ 「 1 3 6 1 6 化 1 ^ 1 0 1 1 , 1『50化5!「6|1 丁 1化00(^6 ^1161+31 0311 1?6 5 3 ^ 3 3113 6 X ^ 0 1 1 6 ( 1

3 . 5

^

1 ) 6 1 1 1 0

3.5.1

3

0 1 1 3 1 1 9 1 1 1 9 州 6 1^635111*61116111: 3 0 3 1 0

0 , 3 乂 3 | 1 3 | 3 | 6 III 3 。 3 * 3

8011 !'06 八50111-06

15 0^31001∼;仆1116 5 八 5 13^16 811^八II丫丁八1^0已丁―8, III 0^031!115 1116 ^ 3 ( 3 50111^6,

1116八(!^!!口!1 1\16^1|31;1 ^ ^ 1 8 0 ^ 01)41011∼35 5016016^1 2.5 0

)

015(313^5 2, 1 4 八 3 1 记 2 . 1 4 8 111 36011011

1116八(,乂;!!^", ^ ^ ( ^ ( ! ^ ^ 八 3 乂 1 8 0 「 0111100 15 3616016015已1116「13「156 1^1;116「

.

5.2 1113^5 1116 11116^31 ^31-1313165 \\ !1|1 1655 111311 2 0 ^15111101 1 ^ 6 1 5 35 ^0111111315 11111055 1卜6 0 1 3 5 8 1 ( ^ 6 1 8 0011111: 7111-651101(1 |3「0|^!1 乂 !5 561 10 501116

01卜6「 111311 20 〔566 0 1 3 ^ 1 3 ^ 2.148 111

!). 10 1卜6 6X31111316

^ 3 1 3 561 81^八II丫丁八11051—8, 1116 ^31-13616 乂八113川1133

19 3151:113

1^61106 5 11 15 1 1 1 3 ^ 6 ^ 3 5 1^011111131 II 13 3 111111161'比乂3!";3616, 3 5 0311 1)6 50611 111 1)10 0 1 3 ^ 1 ^ 5 3.26 111「011^1

3^28^ 丁 0

乂311161 0011111111 111

1116 11163511「61716111: 50^16 0『11115 ^ ^ ! - ^ ! ) ^ 1 0 10161^3!, 「0!10∼

1116 5 1 6 ^ 5 ^ ^ 2 1 1 ^ ^ ^ ^ .

011014 011 1116 13516 1131116

1)3*3 80111*008 111 1116 。 「 ( ^ ^ " 口31化I, 35 5|10\^11 111 015^)13^

3^25^

0)3^137 3.25

固1 9

3 0118^61-38 7; 0 1 0&1& 5 0 ^ 0 6 5 101^6丫丁;^0已!" 8

^!讀1̶67^0̶6 ^-!^] 01991-3013 :

:^ 0111561-8

~"-―

^…、;

"^"

~ \ ::\:;、 ^

― : 「 : ' : : .

^口 :. 1

"八6(56丁 8

;^3^19&165 8謝

;569^116^1 左力6「1131 0313

^11 00^6 耐65

711011 0110^ 011 1116 015(513乂 325^

0011111111 0「11化\^3|1;11)108 ? 1 ^ 0 | 1 0 ^ 3 5 5110\^11 ^ 1116 11101156卩0;1112「 1)1

71115 0(56115 1116

1-1111)168 XVI!!^^^, 35 51】0、乂11 !11 0 1 5 ^ 1 3 ^ 3.26,


0101)1^

3: 1^01-101)16 561^0110?1 0)1(1 ??"〃,、:/^/'〃!"//。〃 。义^0)101)1^

111

015(313^ 3.26

| 10^01

II

2 0

柳 川

^40 ^50 漏 化

口0!^'131 〃0打、化31

^^80 1:则"0卬 11111111 下3【061 【&50

611131^

| ^6^0*1 1^0 ^0 ^0 ^0 ^0 110 ^0 ^0 ^0 ^0

0^

10^611^

~ 0'卬 ^0 110 1*0

、 。――

^0 ^0

^1

11

XV11611 ^ 0 1 1 0 1 1 0 ^ 0 0 1116 31'!~0\\ !11 561601 111(01^01 ( 。 ^ 口 ! 3 乂

0011111111 。。「「亡^口^门^!!门呂10 1110 ^3['!3^16 1131110 \ ^ ^ 1 ^ 3 ^ :

3.27〉 3 1 1 」 0 1 1 0 ^ 0 ^ 1116 111035II 1~61116III 5 0 3 1 6 0^1110 11113111 乂八1131^ 13

011311^6^ & 0 0 1 1*40111111111 1 0 1 0 1 0 1 ^ 3 1 ( [ ^ ^ ( ! ^ 乂 3,28》

013^137 3.27

0x101 柳1 X ^20 卿3∼

,11[!1/1 1叩1/1 1110111

^50 ^ 8 0 1'!00111 1零II

0:0)1

「」

111(6^1 ^0 、011111121 1^0 1^00#^1 ^-! 40 ^0 ^0 6 ^ ^0 ^0 :" '|『1 '!

: : ^

^

^0 ^0 ^0

1- ~

' - 。: ~

― 4 ^ ^

110 1^^

! 0^∼

」 &X1151^

00^

018(313^ 3 . 2 8

― \ 柳 川

柳34 7^40 乂∼50

0 ^

^

1111)111 1(1[!111 116^016 '叩111

111(6^31 "叫",131

^

110 ∼ 011191^ ^1

^

140 ^0

。0^131 職&。

1 &印 140

110 ^0 110 ^0

丁讽

1 、

^

I

^

^ 0 0 ∼

1

」 袖

3『1)11 ^!5011551011 0「 1116 I V [ " ! 1 ( 1 ^ 3 八 ( ' " " ! " 0 ̶ 0 1 1 5 111 01~6311115 3 ^ ^ 1 0 50111X6, 5 6 6 860(1011 2.5 0「

1


112

^^6(110(1^6

1^0^61(11^ 1^7/2 & 4 5 &116"口"!36驢"6": ?1-0011001 80111110128/。卩8143111638

3.5.2

^^1100110178

8 ^ 8 0 0 ^ 1 6 ,。『001111131*1119 6『。11|16|1 。3量690『画。3誦^3酽画3&隱65 ^ 1 1 1 1116 111191*0^1)611 ^31*181)168

10 015|513^ 3.7 乂00 0311 366 \^2^ 1116 0316301*1031

&516 0!\^八1103 1135 11 0&168。『165^ -

丁卜6 0316801*163 0『0!\^八1103 31*6 21*011136(1 10 0 ^ 3 1 6 出 6 ^^!!&1)16 0^0\^^\1103 ^1111 001丫 &11「 0316801^163^ 丫00 030 000113&1*6 1卜6『『60|11600乂 (!&辻比"^。!! 0『1卜6 01*1311131 ^31*1&1)16 ^11)1 1116 ^0\1\3&6 ^31-135165 1131(13 1116 5 八 5 01X16 5|10\^11 III 015^13^ 3.19. 0:8口13乂 3.29 I ^41)16 ^01:03 ! 013311)9 ; | 1^10 ; ― ― ― 一 ― ― . ― ― ― 一 ^ ^ ^ . — 「。1^1。 1183 3&亡8。一化!)。!:!:—。。";

101)16 9一^。1:03 I 111331)10 ;

频 5^3 0x10^

I

| |

0《 1 | 08X081 | \


0 6 0 1 3 : 0 1 1 7 1 * 0 6 1^1011613

611:1||:119

1 0 ^ 6 4 : 0 1 1^65|101130 31111 8 : 化 ―^ 4^ '^^^^; ; 、 .^ I 、1^ ?": 3 二 ' ^ ^ 人 1

4.1

|||1|^||0011011隱"^霧^纏.……翻.^――^―.............^纏"腸…"漏……^ 113

4 . 2 八 I I 0 ^ 6 1 ^ 1 0 ^ 0 1 1116 7 『 6 6 1\11611|0|10|00^匪0 5 ^ 5 6||16^1 !86 1 ( 1 謹 1 1 6 『 ^ 1 1 4 |

4.3

06^6|0|11110111 0 1 1110 7 1 * 6 6 ||| 3 ^ 3 6||16『|1『匪86 1 1 1 1 1 6 ' " 隱 ^ 纏 ^

121

4 . 4 焱 0 6 0 1 8 1 0 1 1 7 1 ^ 0 110||61 1 。 ?『611隱。1 1168|101180 1 。 011*601 ! ^ ! ^ " ! 1 9 ^ 纏 1 4 3 4.5

06^6誦0|1匪119 3 11691*6881011 7 ^ 0 6 1^0116画10

4.6

311111111&!V

4 . 7 焱 匪 X

^

I 。

^

4 ,

^

^

^

^

01110^ ! ! ! ^ ^ 画 " 漏 ^ ^ . . . . . ^ 1 5 9 隱

丁 1^5 0)1&口 16『5110\^5 ^011 110\^ 10 1)1111(1 ^60151011 1^66動^5 10 13(411(1 ?"6^-6851011 1^66動^3 11!"51 6X31111316 5110\^5 1 1 0 ^ 10

^

^

"

^

^

^

^

1

^

6

5

166

10 151*6^101: 3 03(6301*1031 131*^61 30(1 110\^

10 [?!"^]^ & 0 0 0 ^ 0 1 1 0 0 5 1 3 ^ 6 ^ 丁 " 0 6X301)3103 3『6 1^656016(1 丁 116 3

^66動10

| 3 ^ 1 0 1 ^5^0^156

10 6^01 013! 1. 10

6X31111?16 1116 131^61 乂3「1化16 13 6 1 0 3 1 ^ , 献 ^ 0 0 1116 ^ 1 1 1 6 5 7*^5/70/75^ 311(1 770 ^65^01156^ 7116 5 6 0 0 0 ( 1 5

6X31111516 5 1 1 0 ^ 5 110\^ 10 51111(1 3 ! ^ 经 " " 1

13.^1

6 5

0 1 1

6 6

1110(161 10 ^01*60351 & 00111111110113

0 ^ 6 1 1 1136(5 1|1 1116 311(0 1115111*31106 ^ 0 5 1 1 7 , 0細6!乂 1088^6^146170^

1033^6(^146110^

0311 &130 1)6 1110(161601 3 5 3 031680「!031 1:3^61 ^31*131)16

7311165 1)111 ;0 11113 6X31111)16 11: 13 XV&35

3 00011000115 1:3^6^

& ^ & ^ ^ & ^ )

0 6 5 ( ^ 1 3 6 ( 1 111 01、3|316『1》 I I 13^63 0 0 0 0 1 乂 3 卩


114

?^1011^6 ^[。滅"叾XVIII! 8^8 ^6^^186她":?1-00(100180111(10115/01

81481^688

0

^1100110178

丁 116 ^ 0 1 1 0 ^ 1 0 3 10)3103 31*6 00^61*6(1:

9

\^\\&X 18 & 36015100 ^ 6 6 ? 0

9

#

^ 1 1 3 1 13 & ^ 1 5 1 0 0 11*66 0 1 0 ( 1 6 1 ? 戰 ^ &『6 115 0001(30061113? 0

00(16 ( ^ ! ! 出 。 ^

0

口0对6『!0『^。1)31)1111165

0

3831@111116111: 0 『 ^ ! ^ 1 ^ 6 1 3 10 110(^65 3 1 X ^ 6 &1110(16 (^!!&!。!^?

0

\^111011 |3311; 0『1卜6

0

0161110(15『0「 5口11���1118 1116 1^600^5 &X 63011 0 0 ^ 6

0

3616^1113 5^1115

0

111638111*1113 ^ 6 ^01111 0『& 5口111

15 1136(1 &『1)11110111^ 110^6 ^611|11110113?

^1131: 31*6 口0516『1。『&51111163? ^111011 |5311 0『1卜6 6&I& \5 1136(1『0『6511111&1108 ^ 6 3 6

1^01)31)11^65?

9 。 狄 1 0 355180 63011 110(16 0【& 11*66 10 & 131^61 01&35 0

-

9。\^ ^ 0 0 3 6 0 1 6 — 5 6

0 參

^\\&I 31*6 3 1-001: 00(36, 3 0 111(61^116^1310 00(16, 311(1 3 16111111131 00^6/163??

^111011 13311 0 ^

6 ^

15 115601『0『11113?

10 1)11106 0!" 561601: 1)16 ^!^^!-^!^^^ 11*66 。^1131

VI 0『1116

15 1156(1『0『口訓―?

。 ^ 1 1 3 1 01-1161*13 31*6 1136(1『0『

鲁 ?!。" 10 ^ 6 | 0 口 3 11*66 9

1)5108 & ^60131011 1「66 1110^61 10 [ ? ^ ! ^ 『 6 5 | 3 0 0 5 6 10 ^11*601 17131*1(6(1113

攀 0 5 1 0 8 3『6^*6531011 11*66 11\06&] 10 835653 !^!^ I I I 311(0 1113111~81106 鲁

115108 3 『 6 ^ * 6 5 5 1 0 0 11*66 010(161 10 (^乂」。!? 1)1*0^1163 0『「;5化乂 00510016『5 3.1X6 001-50-1*151^ 01151011161*5

^ ( ! ) 1131108 1116 ^6\01^6

1^01)861111165『0『0V6I^-3^I11I)1I|12

0 0 出 6 (:!^口&!"

^6^\(1^x)

4 . 2 焱 I I 0^011^161^ 0 1 1116 7 1 * 6 6 1/161110110109^ 腿II 8^18 6 1 1 量 酽 画 3 6 IV!匪11611

4.2.1

060131011

71*663

八(^(;!^。!! 11*66『6口『656015 3 1116^0111031 36^1116110(1011 0卩出6 63.1&. 丁116 0^1110186^6171 60111*6 6&1& 561, 3.1X6 \1 13 03116^ 1116 7*00^ ?10^6 0『1^16

15出6

丁 116 01*1^11131 3^1110111: 13 11『51 口311 &1011601 01)561^00

111(0 1 ^ 0 0『0101*6 86^11161118 \^ 3^1^1113 & 361*163 0『311111)16 1 * 1 1 0 \ \

1*1)16 33513115 8 0

10 3 56@!1161111)356(1 0 0 出 6 乂3,06 0 『 3 0 1^141 ^6X^)10110101^ "^福^)『0「

1)131 01)361^311011 ^ 1)1 3

&511;011, 630)1 1*65111(1113 5 6 8 0 1 6 0 1 15『1111116『133111110116(1 1010 514^36^6^5 (^^^^\\\\\\\ & 56201601)5 63011 511(3-5631116111: 13 ?1111116「 1?311!6011601 111(0 0101*6 51113-56811161115, 311(1 5 0 011. 丁卜15 |3『00655 0011(1111165 1111(11 0 0 0101*6 口&1^111011;11吕13 口055161巳丁 ^60141-5^6 ^111101\111^

131*00633 0『56^11161111113 ; 03116^

30(1 !1『6511115 111 & 11161&1*01!)' 0『30^111611(3 \^111110 -

03116(1 & 1^66, &1\6 6&。11 36^1116111: 0「 511卜56〇1116111 13 0^116(1 & 1\0^&.

5

丁^16 1 1 1 6 ^ 0 1 1 ^ 13


01X0^)1&V 4: 8"脑"宕060181011

30016110105 1)16 161111 ^6111110^6

7^66 1^0(^8

15 1156(1 10 V 1 0

3 56801601

10 ?^101

0^

/?/^

115

18 13311!1100601 1010 1 ^ 0 0「 010『6

50卜568016015, 丁116 31115-5680161113 31*6 0311601 0^11(1110^68^ ^11611 3 『011|16『,111)60011168 3 1331*601

^68^01186

110(16 1156^15 ^111(1011601

1\0^

八11乂 5 6 8 0 1 6 0 1 0「 50卜562111601 11131 13 &11116『1)3111(10116(1 111(0 010『6 5116-568016115 0311 &150 6 6 03116(1 &11 1(8 8110063301*3『01*015 3 1)^0110^1 0『1)16 1『66^ 丁[!6『1031

& 0 117161'!116(11016 "(^.八0036

56^11161115 11131 31*6 0 0 1 | 3 3 1 1 1 1 1 0 0 6 0 1 3 1 * 6 0^116(1 /^/!!/""/ 110(^68 0『1&0/110^65 0『160^68 0^1*16 ^3011 163^00016 18 3611116(1 \ ^ 3 11111^116 0011151113(1011 0『1^11363 0? 1)16 乂3|1165 0 ^ 1116 111)3111: V3I^I3^)1^5. 7116 163^ 110(163 31*6 ^15^01111: 3111)56(8 0『1116 01^1@11131 6&X&. 丁 出601, 31\6 63011『6001^ 111 1116 50016110163『6化『10 3 163^ 110(16 35 3

13 0 0 0^61*13^ 1)6(^6611

0 0 ! ) ' 0 0 6 163^ 1\0^&. 10 11115 ^ 0 0 、 I

3611)610085 10 0 0 6 ^014^

0 6 0 8 3 1 0 1 1 7 1 * 6 6 1^10^618

4.2.2

八016(^51011 11*66 1110111^63 111(61*1116(11316 3 5 \^611 38 16111111131 110(163 ^60181011 ^66

!!^^). I I I 000115仁1116

乂011 (^[^。乂卩0『1^6(11011011 0 『 1 1 1 0 1 1 1 ( 1 6 3

1710(^1,

0 0 1 丫 1116 163^

加 ^ .

八 3 6 ( ^ 5 1 0 0 11*66 010(161 13 ( ^ ^ ( ^ ( ! 0『56乂61^1 口3115, 311011 &5

,110016 361111&10115, 0『「11165, 10 3331^11 63011 1*6001X1 0『& 6&X& 5 6 1 1 0 3 !63『110(16 ^

^05(6^10^

^01)01)1111168

0『63011 163^ 00(16

1116 88513111116111: 0『3 10^611^6110

63011 163^ 110^6

1^0爿6 016(1111(10113 3『6 416^610(5601 1131113出 / / " " / " / ^ ^0(0

861 30(1 31*6 3(3(6(1 111 160115 0『111(3111:

6

?05^101~

^01)01)1111168 31*6 03101113(68『0『630)1 00(16 1131113 1116 //^/"//^ ^

1)16

1 ^ 6 1 10 630)1 110(16 !3 &130 010116 ^1111*1113 ^

?0516『!0『1)1*01)31)111(163 &1*6 0^)861^6(1 ^0^110118 ( ! & 协 7 3 ^ 6 1116 6X3(111316 0『& 6 1 0 8 1 7 乜 ! ^

6 1

1 6

^!"!^

0

861. 丁116 385180016111 0『

^11356 1131113 1116 1^(1111111^ ^010 861.

0『131^61 1 ^ 6 1 5 XV111110 63011 00(16 I I I 1)16 11*31111118 .八^031^

&

3

5

16^615 ^111011 0 2 0 1)6

1-6(31*63611(6(1 6乂 7*^5/70/156 31\6 110 ^651?01186 0!" 1 3 0 3 0^ 丁卜61)0316^10^ /?"0^6/7/7^ 0『『65(30056 111 3 9

110^0 13 1116 ^ 0 ^ 0 1 1 1 0 1 1 0^『6001X15 "!出1116 51111113「1 乂,

6 ^05^10^

^01)0^111^

&I

60)031 10『65口0056 0 『 1 ^111110

00016.

0『00『65口0056 0『3 00(16 15 1116 |31^0130111011 0『!"6001X15

131-^61 16乂61 6^1131 10 0 0 『 6 5 口 0 0 5 6 , 0 『 0 , ^1(11111 1^X3.1 110016. 711656 ^08^10^

^01)01)11^165

^611111116(1 ^111*1113 1^16 ^1111113 0 ^ 1116 11*66 & 0 3 1)16乂 1)600016 1)311: 0 ^ 1116 ^ ! ^ ! ^ ! ! ^ 6 6 111611(10116(1 &1?0乂6,出6乂 &1*6 08101113(6(1 1151113出

丁 116 3 8 5 1 @ 腦 0 『 &

6

1(16

31*6

^(!^!.八5

( ! ^ .

16乂61 10 &IX 11)^1011131 「600!^, 0「 10 & 110(16 3 8 & \\^016, 15

8

^60151011^ 丁 116 (^(;!^。!^ &1*6 111&^6 111 01^6『10 1113X1111126 |3『0^11, 01101111126 0051, 0『1111111111126 1X115013551^103(1011 61^0『,丁0 !11115(1^(6 & 36(^51011 1)35601 0 0 ^0^\1 6X81111316 0『& 1)1031^ X 1 . 。111750565, I ^ ^ ^ ( !

1113X111112&11011 (^!!&^^! 1110

I I I 01^6『10 1113X1111126 |3!"。1115, 乂011 1166(1 3 ^1*0^11:

1116 口 『 0 犷 " 8 ^ 6 0

1(1 731)16 4^ 1 ^

1

?0『II 1 1 1 5 ^ 1 ^ 0


116

?1-6^10(1^6 140(^61111^ ^///? & 4 5 ^16^1-186滅"6广?^001100\《。/"//膽^

81431^7655 ^^1100110^18

731)16 4.1 1)60181001

1)601810112

310

0 0

16^61/01388 1 0

一 51

10 丁&616 4.1, 0 6 0 1 3 1 0 0 1 11163113 3531@111113 3『6001X1 10 131^61 16乂61 1 (!^^^^!^匀,311(1 0 6 0 1 3 1 0 0 2 016&05 8351^111113 & 化 。 。 ! ^ 10 1^&X \^\^\ 0 01011-『65|30056丫 1)113 ^0^\1 01&汁1乂 111(1103163

1『3

11116 !^?口。!^" 15 001X6011^ 013851^16^ &5 &『65(300(16!", 111611 1116 ^1*0^11: 6&1116^ 13 810. 1^ 3 11116 1100『^!^。!^" 15 013531116(1 3 5 &『65;30036『,& 1085 0 ^ 8 1 15 ^ ^ ^ ( ! .

1『& 11116 !^^口。!^^ 15 018551^16(1 3 3 3

(!^!!-^^^^!!^^!", 11160仇6 ^ 0 ^ 1 1 15 26『0、 1『&忱116 !^!!-!^^^^!!^^!" 13 013331116(1 3 5 3 !!^!!-^^^^!!^^!, -

1)1611 1卜6 |?1^0111: 15

261-0.

51^(3056丑、1*6^1*656015 1116 6X13601601 ^0^\1 \1^661 0601310111

9

8.1X6 ^

2

^6^861115

1116 6X^)601601 131*0111

110016『0601310112. 1^丑、〉五 , 11160 0 6 0 1 3 1 0 0 1 13 13^611 & 0 ( 1 出 6 1*6001X1 13 3831^116^ 1 0 出 6 131^61 2

1^61 1 . 1 ^ ^

2

^ ^ , , 11160 060151011 2 13拉化611, &0011116『6001^ 13 3331^1160110 1116 1 & 1 6 ^ 6 1 0,

0 ^ 6 0 & ^0^11 0131^乂, 『6001X1

&11(1 ^

丁 116 6X^)6016(1 口 2

2

3『6 1)356(1 0 0 出 6 ^05^10^

^01)0^1111165

0『1116 110(16 10 ^111011 1卜6

10 ^ 6 1 6 ^ 1 63011 !"6001X1 13 3531^116(1 1116 ;30516『10!" 1)1*01)31)111(165 0『1卜6 00(^6 111 \^|110|1

11 30(1五

5

^ 0#力十0

916

5、 &1\6 ^

户 『0『63011 1^600^ 0

5

2

〉 31*6 0310111^6(1 3 5 丑 、 : 1 0 # 《 一 16「6 ^

311(1 ? 1*6口『65601出6 口0516『10『^0^6111(163 0卩 0

1-631301136 311(1 00『65|30056卩0『1116『600『丄

I I ^6110\^3 & 0 0 1 1116 ^0^\1 6^)113(10113 9^0\6 1卜&I &11 1卜6 [^(^『(^ 111 3 00(16 1 1 ^ 6 1^16 53016 150516『10「 口「0153131111165, 31\6 31*6 1116『6『0「6 3551^116^ 10 1116 531116 \31^1

&].

八0 0X3011316 0 『 & 1『66 !5 3110^11 I I I ? 4 . 1 . 丁 116 11*66 5110\^11 111 I 10 111^111

7

「 6

4.1 13 1 1 3 0 ( ^ 0 『 & 代 1 ( 1 01^6『

1116 131-00658 0『&3513)11113 131*861 16乂615 10 1卜6 !!^^.

丁1161*6 !^!^! 1)6 3 1055 ^1160 & 11116化^!^!^^! 15 01358106(1 35 3 !^!!-!^?口。"!^!. !"10\\^乂6「, 1(1 11113 6X^I11[^I^ I &01 &83111111113 ^ ^!!^!)!^!^. 7)115 38511111|^1011 0311 1)6 1^13X6(1, 30(1 &11也6 ^ ^ ! ! ^ 8110^0 1101^ ^01)1(1 3(5151乂. 1

1

261*0『0!"

1116 (^(:&!。!! 11*66 5110^0 111 I&11『65 4.1, 4.3, 311^ 4.4 !5 ^6\6\0^6^ 1151113 1116 060151011 丁1*66 00^6^ 90\^6乂6『,1^16 11*06 ^13^1^5 5|10\^0 31*6 11311(10^0(1『0『!11115(1*3(1113 乂3「!0115 5(6(35 0『0)0 11*06 ^610(31116111151-00655. 7


4: 81111(^111^ 0^01810^1 7^66 1^0(^3

I。 ?^65^0(136

0^

1^18^

117

^19111*6 4.1 1^01)6 1: 1^001 110116: ^ ^ " ( ! 化 ^ 708, "。!!-^,&!^ : 6444, 丁。131 : 7152. ?I ^ 0.099, ?0二 0.901 , ^001 110^ ^601510111 : 0.089, ?1-00! 11!]661 ^60181002 ^ 0 0601810111

,'

525々

1^0116 2 1^)00(16^5 : 251 , !^-!^^!!^巧二 2868, 丁0131 = 3119,1*1 二 0.08 ,?0二 0.92 ^ ?1-001 ""(^!" ^60151001 ^鲁。.12, ?「0「|111111 (6^ 36(^510112 ^ 0^ 000|8:0112

1^0(16 3 ^651)0110161^ ^ 457, !^-!"。卩。"(!饥^ 3576, 70131 ^ 4033 1*1 : 0.113 , ?0:0脱0(11110^(1601510111 ^0243^ ?「001 !!!^^ 3601510112 ^ 0^ 1)601310111

丄 06(111 5001-0

八86

,' 540 35

540

35

70131 2 1438 ?] ^ 0.07^ ?0 ^ ^93 000151011 1^「0「11 ^0.23 0

1X0110 5 ^ 1^31^2: ^011|12》 1(6513011(161^ ^ 151 1^00-1^51500(1613 ^ 1530 丁。131 = 1681 ?] 409,?0、91 060151011 ?1-001 I -0.01 2 0

1^0116 6 ^ 1^31*3: 8『0叩3》 ^5^110161^ : 226 ∼。?!-!^!^)!^!"5 : 1944 丁0131 = 2170

06015:011 2

1)6^1810112

0(01810111

1^0(16 4 (!^! 1: 81*0叩,) 11651)00(161^ 100 1

3

∼。!!-!^网^。 : 1338

060151011 1 2

1^0116 7 (^! 4: 81*0叩4 ^ 116515011(16^5 = 231 !^-!^^!^。 : 1632 70131 ^ 1863 1

?「0「11 0.144 0

?1-001 0.364 0

0001&1011 I 2 060(810111

?]2111*6 4 . 1 !3 3 0 6X81111?16 0『3 11*66『0「 1116 131^61 乂 3 『 1 3 6 1 6 丁 11!5 V31*13616 1135 1\^0 16^^15: 1 『0『1*65^)01156 &1\6 0『0「 00『651301156. ^3011 0 0 ^ 6 8 ^ 0 5 1116 10『0『01&11(30 0 0 1116 ! ! 歸 ! ^ 0『1*65^000101*5 311(1 !^!!-!^!^&!^, 1116 |3『0「011100 0『1^65|5011^6^5 (?!) 311^ ! ! 。 ! ! - ! ^ ^ ! ^ ! ^ ( ? 。 ) , 1116 ^60151011 「68迂『出"8出 311003(1011 0『& X&\&\0

6&0^1 ^ ^ ^ ! ! 『 0 『 1 1 1 3 1

00(^6, ^1\6 1\\& ^0^11

6

门^^^. 3 6 0 1 1 0 0 4 . 3 . 4 2 1 ^ 0 5 & 516)3-6^-51615 111113(1*311011 0 『 卜 0 ∼ 『 ; 5 6 ^106『36「1乂65 1110 10^01111&1100 8 ^ 6 0 111 63011 0 0 ^ 6 0『1116 IV&&.

4.2.3

0 6 0 8 3 1 0 1 1 7 1 ^ 6 11/1011613黎5, 1.0913^90 1 ) 6 9 ^ 5 3 8 0 1 1

1^0^613

八(^。^。!! 11*66 0 1 0 ( ^ 1 15 000115056(1 0『3 561 0『『1*165 ^ 1 0 1 1 0 3 0 1)6 3 ^ 1 1 6 ( 5 10 1?311!110!1出6 ^ 3 1 3 1010 ^18)01111: 31*011135. 110111^6 1116 10313(10 1^6^!"6551011 0100161,出6 1『66 11100161 ^ 0 6 5 0 0 1 113^6 &0乂 60111311005 0『(^^!^化^!!佑.105163(1, 11: 00013505,『0『630)1出5』0101 ^01113 0 『 1 & 0 / " ( ^ , |30516『!0『口「01)3131111165 \^111011 31-0 111601561^65 1136^ 3 3 ^1*6^!0^6(1 ^ ^ ! ) ! ! ! ! ^ . 711656 ^05^10?'

^01)0^1111165

3『6 ^ ^ 0 1 0 1 3 6 ( 1

1116 1^01111^ 0『1(16

4.2.4

^ | 1 | | | ^ | | 1 9 1116 0 6 0 1 3 1 0 1 1 了『66 1\00|161 1 。 ^ 1 * 0 3 1 1 6 0 1

丁 116 5 八 5 00(16 ^ 6 ^ 6 1 ^ ( 6 ^ ^ 131-6(11011011 30(5

1^16 丁1*66 00(16 0311 1?6 ^ ( ? ! ^ 10 1110 ( ^ ( ^ ( ^ "

5&I『0『口11「口0565 0卩

丁 116 0 0 ^ 6 口13065 630^1『6001X1 I I I 0 0 6 0『1116 131-60(6(106(1

丁116 131*6(110(6(1 1^01)31)1111^ 0『& 1 3 ^ 6 1

^"(^

/?^^^///^ 3530013(6(1 ^11)1 1116 160/110(16 1010

^0^?5\

38 ^^^^!!^^)『0「 63011『6001X1 15 1116 口0316?"!0!" 出 6 1^600^ 13 ^!^^^. 811106 & 13X^1

1 ^ 6 1 !5


118

?^1^6

^0^611^

^///? 848 ^16^156滅贈:^00(100180114(10125^

81431)1688 ^^1100110118

3381806(1 10 6&0^1 163^00(16, &11『6001X15 &11 1010 & \63^ 00^6 &『6 333180601也6 53016 ^ ! " ^ ! ^ ! . ?0『6乂細口16, 1116 131-^61 1^61 1)6『65口00016「 0『000-『651300016『,

0310111318011 。11116 1/^01111 0 1 3 7 『 6 6

4.2.5

71161-6 &!"6 311113110113 I I I

11 13 0 6 0 6 3 3 &

030 1)0 0^10111^601 115108出 ^311(1^1011

0 ^ 0 1 1 1 ^ 6 1)16 \^0^

1 0

& 56仁1651:

6

0 『 & IV&&. 丁卜6 ^01111 0『& 1「66

& 361, 0『30^ 0也6『^313 561 ^1161*6出6 1 ^ 6 1

31*6 1 ^ 0 0 ^ 0 &1\6 1116 |11|31115 0 6 0 6 3 3 3 1 7 『 『 ^ ^ " ! ^ ^ 6

^0^65 31*6 ^&!!&!)^. 0 | 1 1 ^ 163^

0

1100163 31*0 115601『0!" 03101113(1113出6 "。(!卜.

10 ^0101131-186 ^1106『,1116 ^01111 0【& 11*66 13 03101113(6(1 ^ 1151113 1卜6 ^311(131100 ^3(3 56110 0001|33^6 11*065 0^ 011^61-6111; 31263 111 0『016『10 出6 11*66 ^1111 1116 0)3(111111111 !^!!!!^!" 0^ 16^05^ 丁卜6 \^011|1 0『 & 11-60 0311 &150 1)6 03101113(0^ 1131113 ^ 『36『10 0001(531*6 1116 1)61^011113006 0^ 1116 ^6013!0111『66 1110(161 01116『1110^6|5^ 10 1)0111 1(16 03565, 1116 1116(110(1 0^ 1116 ^01111 0^ & 111

1 6 5 1

0

11-66 1 3 出 6 3&1116. 1-161-6 I 111115(1*3(6 1)16 001X1(311(3(1011 0『1116 \^01111 0? & 11*66 1131113 1116

1(13(1011

& 361. I 1136 3 0

010(161 ^1111 3 1)111317 ^ 8 ^ 丁^ ^102117 II&3 1^0 16^615 0『01&5565. 7116^ 31*6 1*631)01136 (!!!^^^^ ^ 1 〉 311(1 00『65口0056 ^ 111(11021(6(1 6^ 0 扒 I 1136 口『0化33 & 11163311『6 0^ ^01111, 3X16 8\\0\^ \\0^ ^0^\15 0『1)16 163^ 110(163 0『3 11*66 31*6 0310111^6(1 1卜6 0『0111 0131「1乂

6X3171(316 0【& 1*63(301136

6

16

111(1*0(11106(1 111 丁&1)16 4^ 1 , 丁 116 0310111^(1011 !5 3 1^0-510(3 ^ ^ ^ ^ ^ ^ . 10 516|3 1, 63011 !"6001"(!『『001 1116 V』化31!011 6&1& 561 13 &331^116(110 & 11036 ^6^1111110115出31 31*6 ^6^610136(1 (^『!!!呂1116 11^1111113 口 口 1306(1 111 &

『0『11131:

110(^6

1)386(1 00 111165 0『

八5 110(0(1 31x^6, &11 1*6001X15 11131 31*6

13311101|13「 163『00(16 &1*6 335&116(1 1116 531116 |30516『10『(?^^!)II 1(165 ^ \ \\^6『6 001X1(311(6(1 110(^6 ^111*1113 1卜6

|311356 115!!^出6 ^1111113

361. 8!111113『1^, &11 1*6001x15出31

& I I 111(0 & |531^0013『163^ 00(16 31*6 3 5 5 ^ 0 6 3 1卜6 331116 131*801 01833/16^61

(巧^!^!!^^!^

0『0011-

11131 ^ 3 5 0 1 6 ( 6 1 1 1 1 1 1 1 6 ^ 『 0 『 \ ^ 110(16 ^111*1113

10 5160 2, 1116 1)1*0111 13 03101113(6(1卩0『63011 I&110(^6 111 6 & 0 扒 1 ^ 6 0 0 ^ 0 ^ 1116

13^61

1(13(1011

0『1116 ^66 1)386(1 00 1116 00(1401 \^\\\& 0『1卜6

5&1.

?10(^6 13 01333111^ 33 3 !"^「(^(!" 0 0 3 6 (^^!^^^ 1116 13『8 ! ^ " 尹1 〉, 11^1113

1『3

61

1110 0011101

1^61 0『1 3 0 3

11131 163?00^6 !5巧# X I 0 十 『65口0036『00(^6,

#

6

0

『6001X15

1^600^3 ^11111110 0011101 ^^61 1^61 0^ 0, 111611 1卜6 |3『0化0『

( 一 5 1 〉 , 1『,00 1116

1130(1, \\\& \&3^ 110(16 15 013551^16(1 35 & 0011-

11160 1)16 口『0厂11 0『出&I 00016 1 3 巧 孝 5 0 十 / 1 。 5 0 ^ ?0110^102 11113 151*006(1111*6, 1116 #

卩『0111 0^ 630(1 163^ 00^6 0『1116 ^66 13 03(0111816(1 3X1(111160 3111111116(110 0^101113(6出6 10X3.1 ^0^\1 0『 1116 I!"66, 丁116 ^6?'0^6^0^11

0『1)16 11*66 13 1116 10^1 ^10^\\ ^ ^ ! ^ \^

10(^1 II(!!^^!" 0『1*6001^5 1|1

1:『66,

八5 111611(10116(1 1?6?0『6, 1)16 10^1 ^1*0^1 311(1 3^61*3^6 1)1*0111 0^ 3 ^60131011 1「66 1110(161 0311 &150 1)6 03101118(6(1 1131113 ( ! ^ 561. 711!3 13 1101711211 !)' 30116 乂011 ^301: 10 00171^31*6 ^61^001131106 0『& 0160;51011 11*66 11100)61 10 0出6!" I I I 6(1161*131*136 \11|16『1116 ^1*0111: 11131『^ 0311 1)6 611161*6(1 ^11611 乂011 01*6316 3 ^ 3 ( 3 3011『06, 3 5 5110^11 !11 0)^2^&X

2,

013口1^5 2.15^ 2.16, 2.17, 2^ 2.19. I I 0311 &130 56 60161-6(1 \^ 56160(1113 1(16 1110111 1 ) 3 ( 8

0 0 3 6 311(1 0110^1113 00 口 ^61^11(8 (^!), ^011 0311 1116 口 『 0 『 1 1 腿 ! ! ^ .

1003^6(1 10 1116 \\^\ 0^ 1116 1)601810118 ^ 0 ^ 6 ^ ^ 8 ^ 56160(1113出 060181011 6

1116 131*0111 11\3^\\^

4, 1 8110\^8 1卜6 ^10010^ 111 ^111011 ^011 6016「


4: 8111!^!^ 0(^0151011 丁"^0^6^

101 ^^0(1X6

011^ /?/^^

119

广^11一;16

^3X11^116 1^61

10

05030^!"7 0501510^2 00 -1.0 !0.0

… 10.0

1 0~

0^

46驷

013^13^ 4.1 3!50 5110\^5 11131 1 113^6 56(6016(1【I化^60151011 ^111101)011 1\13\111112^ 丁 1化116X1 516^ 15 10 1611 1116 1)00181011 71*66 1101^6 10 1156 1116 口「011【11131 「;X 011(1 1116 ^60151011 1^111011011 I 5 6 1 6 0 ( 6 ^ ^ 丁0 ^ 0 III!5 I 561601 1116 1)60181011 ?

110^6, 311(1 111611 111」6「 8111)1:1*06 111 1116 |3「0|^11165 ^31161 0!" 11化

1)60181011 1^00 110(^0 ^5110^11 111 0 ) 5 ^ 1 3 ^ 4.2》,I 561 1116 1^10(110(1 ^ ^ ^ ^ ^ 10八556551116111 311^ 1116 ^\356551110111 ^ 6 3 5 1 1 1 * 6 ^!"0^)6(1^ 10 000151011. 「;56 1^1106「, |31~0^|1 15 1101【I化011卜0^1^[-1011 11131 0311 1?6 115601 「0「 0 3 1 0 1 ) 1 3 1 1 1 1 5 出 则 ^

1)1

6

0卩3

丁|16「6 31-6 3 11111711)6!" 0 ? 3116111316 1;「116「;3, 丁 1化3^3113616 0「116「13 0「 11163511「65 0311 1?0 5 6 0 1 13^ 0110^1115 I I I 1116 ^31116 0011111111 0「 1116 ^886881116111 1^163811 !'6

11115 1 ) 1 0 ^ 5 11(3 3 1116011 0 「

3116『1131~6 3556551116111: 17163511「65, 35 5卜0、^11 111 0151)13丫 4.3,

0:3|3|37 4 . 2 ^ 融 6 6 關 刚

1^111111)6!' 0【匕63^63

| ^33633111601 1 060181011 0.25

013^13^ 4.3 ^35885^80111/1835曙 ^5565 51716^1 「「3011。门

060151011

日0^6【「0111八1^115101.80—1一. ^ ^ ^ ^ 1(103)1011 711718 0 ^ 3 5 8 ^01)051016111 1^

八6「|6「^650['! 1311011 0「03011 11163511「6 0「 1111X161 ^61*^0['!1131100 15 ^ ^ 6 1 1 ! ^ ! 。 " . 卩 0 「 「11111化「 ( ^ ( ^ ^ , 1 0 0 ^ !11 1112巳1116「卩「!56 VI"化「卜"卩『3」1;1:》


120

?^101^6

140^11^

801^10^5^0?' 8148117655

848 ^16^186她":?^001\00\

^1x00110^18

1)60181011 1『3 131-0111;

15 60161-6(1 311(1出6 36015100 ^1111011011 ^3X1111126 13 3616016(1

1)1611 601:6『|3「156 \1!06『03101113(65 3 乂 6 『 3 8

6

3

8

10

4.1〉,

^ 0163501*6 0『1110 ^01111 0『3 1^66 311(1 3616013 3

11-66 5 0 &5 10 013X1111126 11131: \3:\\1&. I? & 0051: !^&!^乂 13 60161*6(1

1^16 016(^5100 !110011011 ^101111126

15 5616016(1, 11160 50161^^156 ^1!06『011011111265 0 0 3 1 10 561601105 3打66, ^ 6 1 ^ ^ 6 8^1131*6(1 0『0『 丁卜15 13 1116 3乂6『38 『 ^ 1 1 2 1 1 * 6 0 ^ 1116 ^1^61-61106 1 ) 6 ^ 6 6 0 1116 ^ ^ ^ ! ^ ^ 0 0 & 0 0 1 6 30011^16 301:1131 6 0

011100016. 丁卜15 016&501*6 13 010「6 3口!3『0|3『1&16 \\^160 1卜6 131*^61; 18 000110110115.

1^18013881^10311011 115108 1116 6X31111)16 0『&『65(30056 010(161, 1『1116 1110(161 01&5311163 &『65口00016『35 & 11011-I^63I30I1^1^I^ 0『 0135511165 & 000-『65!300016『33 3 1*651)011(16!', 111611 1)16『65011 13 & 1111501355111021:1011^ 7116 11113013331^0311011 1 ^ 6 13 10^61*36!)'『61&16(110 1卜6 ^01111 0 『 & IV&&. 861601108 11115 0口1!00

103111101:

6016113「156 1^106『10 561601: 3 1 ^ 6 5 0 3 8 10 1111111111!26 1卜6 1111111|36『0『

10 1116 0 3 3 6 0『& 1*65(30056 010(161, ^111011 1135 & 5 1 0 & 1 : 3 『 8 仁 ^ 6

1116 30111&I 1 * 6 3 1 3 0 1 1 5 6 0 『 1 ) 1 6 1 0 ^ 0乂6『311 1-631)01136 1*316 0『1卜6

"!!^" 15 0310111^6(1 3 5 1116 1*31:10 0『

0 『 1 ( 1 0 腿 1 ( 6 ( 1 0 1 ) 5 6 1 ^ 1 0 0 5 10 1116 ^ 1 1 ( ^ 0 1 1

1013(1011 6

1*651301136『0『6&011『6001X1 !0 1116

1 6

&

801: 1 。 1116

116 1*^111(1113 13 ^0116 6乂 1116 ^1*6011016^ ^01)31)1 I I ! ) ' 0 ^

1(1&1100 6&X& 861.

丁 116 136^0611138 ?『301^(1 6 & ( " ^ ) 10 1)6 1156(1 & 5 出 6 133513【0『1116 1 6 & 0^101113(1011 13 3(3601^6(1 6乂 6 0

361(1113 1116八8868801601 ? 1^0(100 ^0^61^ 0.25

^

0

4.2.6

0?出6 1)60181011 7^66 。 ( ^ ^ . 丁卜6 ^ 6 ^ 1 1 1 1 乂3106『0『11115 !5

/ 。 ) .

3 。 瞧 6 5 0 1 1 1 1 6 71*38111119 31111 \^3隱!||31謹011 0 3 1 3 811 1116 。6^6画001116111 0 1 3 060881011 7『66

丁0 36乂610|3 3 1『66 1110(161, 乂00 1166(1 1 ^ 0 6&X& 5615. 丁116『11*51 15『0『11*31111113 1(16 1710(161, 311^ 1116 3600001 13『0『^1X1111113 0『^106-111111118 ^ 6 0 1 0 ^ 6 ^ 八 ! ^ ! ^ ! 6 ^

3&1 18 0(3(1011811^ 1160(16(1『0『31\

&^&1\6&1\1 355638016111: 0 ^ 出 6 010(161 , 711636出『66 6&1& 36(8 &1*6 !"6^51!"6(1 10 !0 ^01611)1*136 ^4106『35 乂311331;011, &0(1 1651,『65|56011乂61》丫011 0311 3 6 0 ^ 6

^61-06111326 0『1116 1110^61

561

13 10 1)6 31100^(6(1 10 63011 0 ^ 111636 1|11^66 13111150363. \\^116 11161*6 8『6 0 0 1101^01111 !"11163 ?0『&11003^011, 乂011 0 3 0 &110∼061^310 114163 0『111110115. 1『1卜6 6^& ^\^\\3^\6『0『1110(1611113 13 13『〇6, 乂011 0 3 0 11 60)11311 乂 1 ) 6 ^ 6 6 0 1卜6 1111*66 6&1& 8&15. 1『1\\&I 13 0011卜6 0356, 1136 3011161:111113 111(6 4 0 ^ , 3 0 ^ , 311^ 30 石,0『500/0, 25 入,31101 250/0, 丁卜!5 13311111001118 15 60116 ^ 0

0

5^)601^108 ^

1

6

^^11163『0『出6 7|^!0|118,

丁6811)1*0(3611165 0 ^ 1)16 1)313 ?31*11(100 1106&. 116561^108 0101*6 ^3(3『0『出6 11*31111113

\^1!|1311011,

63.1& 561: 86061*311 乂 1*6511113 111 010『6 5(31)16 口&『301616!"

4.2.6.1

丁『3!11||19| 0 3 ( 3

丁 116 1^0111111^

8&I 18 1156(1 10 口61^0『01 1111*66 1113111 ( ^ ^ :

鲁 ( ^ ^ 。 口 ― 1*11163 10 3551311 ^ 6 0 0 ^ 3 10 110(163 ^1X0(1& (^""!。"!5、 9

03101113(1113

|30516『10『131*0636111(163 (^^ |51^0|30111011 0『08863 01 1*6001^5 111 63011 1^61 0『

1116 13『8 0『。『6^011 110(^6 6

-


。―化"4:

7^66 !^0^13 10

8"脑"宕0601510^1

101 ^6^01756

011(1 ^18^

121

4.2.6.2 7311013(1011 0 3 ( 3 丁116 \011^01101\ (!&13 561 13 113601 10 ^14^6 1116 XV&&, 1.6., 3 6 1 6 0 1 出 6 ^1^1-8126^1 丁 116 1011131 11-66 13 11311311^ 1 3 ^ 6 ^ 11113 15 03116(1 1卜6 1)10X117101 (^.

^66 0「 。 / ? / / 廳 7 1^66^

8 0 0 1 6 0311 I I 1116 \\^。!6 ^66^ \\\ 1 ^ 5

6 0 0 化 , I 1156 111636 1\^0 160115

8乂 1-6010^108 06113111 51*3001165 0『1卜6 1)10X1111011^6I 出1?6「6111151*30011, I ^ ^001110^2^

0311 01*6316 501&116!" 11^65, 8乂『6010乂|02 3 0卩111656 11*665 1135 0 0 1 乂 0 0 6

& 011^51*61111『66^ 7116

31101 1卜6 131^651 15 1卜6 1111(131 1710X11^701 IV&&,

^1^01-6111 6^3001165 ^1-0(111063 3 0001136『0『^05511310

0311

1 6 3 卩 1 1 1 6

0130乂 16^65^

丁 11115 「 6 0 1 0 ^ 0 2

51115-1『665, I I 15 060653317【。561601 0 0 6 0『1116111 ^

&1100 6&1& 13 11563『0『1)16 56160(1011 0 『 3 511卜1「6匕 八5 6X^131116(1 10 86011011 4.2.5, 1116 \^011|1 0 『 1 1 * 0 6 0 ^ & 1^3(1011 『6001X15 0 ^ 1116

1(1311011

561, ^0^\1 15 V

1

0 3 5 1)16 \01\^01\011 ^0^11 111巳!!&!^!^^ 1^106^

^311(131100 131-0111 15 1156(1 10 5616011卜6 0^1171101 ^\^\1~3\7&(^) !!"!!!!^!" 0 ^

! ^ " 5 1 2 6 13 03)01113(6(1 0 5 1 1 ^ 1卜6

56伫311(1 0116 0『1116 01635111*65 0『^01111 15 「「0厂11 311106 口「0^1 15 03101(13(6^ 1151113

111 1116 1『66,八0 0^賺I

3 0 ^ 51113116『11*66 (& XV&&

^66 \5 635出31汁66

1「66^ 丁 116 5126 0 『 3 11*06 13 01611116^ 3 5 1110 11131 ^101^3 1118116「 口「0^11111311

『6∼6『!^^^^^) 311^ 6(^1131 0『^!^!!^!" 131*0111: 11130 3 0 ^ ^ 6 6 11)3.1 \5 13^6!".

311106 I I !5 635乂 10 11*3^1 1x0111 110(^6 10 00016 311(1 0011511*110111*065 1\\&I 6 1 1 ( 1 叩 ^ ! ^ , 53乂, 160 163乂65, 1)161*6 171

1)6 1110『6 11130 0 0 6 11*66 0 ^ 3 8 ^ 6 1 1 5126 ^1(11111出6 013X10131 〈 0 『 ^ ^ ^ ) IV&&. 50101^51^156

1^1106『5631^1165 111「00^1 &11 0『1116 ^05511)16 1『665 10『10(1 1116 0 0 6 ^1111 1(16 11^11651

1(13(1011 |5「0厂11.

0 6 ( 3 1 1 5 0『11115 3 6 3 ^ 1 1 |3『00655 31*6 0113011356(1 1口 3601100 4.3,

4.2.6.3 丁681 0313 7116 1651 6&X& 5&I 15 1156(1『0『311 10(16^)611016111 355055010111 0『1116 1111&1 010(161, !^!"^。" !3「1 乂 \^|160 ^011 \ \ ^ 0 1 1 0 0001|531^6 1卜6 |361^0111131106 0 ^ & 016015100 11*66 010(^61 10 1116 ^61*^0111131106 0『01116『!!!。^^^.

4.2.7

1169^331011

丁『66

\^1160 1116 131*561 13 000110110113 1116 1「66 15 0 3 1 1 ^ & 1-6^6881011 11-66. 丁 116 ^ 6 ^ 6 5 5 1 0 1 1 11*66 010(161 5

0011515(5 0卩1卜6 111163

( ! ^ 门 ; ^ ! ^ ) 『 0 「 口 3 池 1 0 0 ^ 1卜6 0 ^ 3 501: 1010

00(165 30^1 3130 1116

1116311 0 ^ 1116 13^61:『0『630)1 163^ 1\06&.

4 . 3

0 0 ^ 6 | 0 | 1 1 1 1 0 1 1 1 0 11 1 1 67 1 * 6 0

I I I8

^

8 611101*111-830

11/11116酽

00115111101100 0 『 3 36(^51011 ^ 6 6 0『1*681*6551011 1^66 1|1乂01化5 1 ^ 0 013」0『516口5, 丁116 ^!"51 516口 15 81*0^108 1116 I!"66,八5 016011006(1 1?6『0『6, 0130乂 51113116「 1『665 (^^^-!^^^) 0 『 5 1 2 0 5 ^ 6 01*63(6(1 6乂『61X10^118 0 0 6 0『11101*6 1)1-31101163 0 ^ 1116 1 3 ^ 6

0 3 0 11160

丁卜6 56000(1 516卩15 361601108 ^ 6

0)3(111131 511154『61 80161731*186 !^!;06「 0泡『5 & 0 0 0 1 1 5 6 『 0 『 0 1 1 0 1 0 6 3 犷 0 『 ^ 6 1 ^ 0 1 1 1 1 1 0 5 丁 116 ^0110^1113 360(10113 5 1 1 0 ^ 1 1 0 ^ 10 1136 111656 0161:110013 6乂 36111118 乂3『10115 |31^0[?611165.

4.3.1

6 1 X 1 ^ 1 1 1 9 311墜118^31 1*1*66

八 5 016301*11)601 I I I 8 6 0 1 1 0 0 4.1.1 , ^0\^1115 3 11*66 !0乂0卜65 511006331^61^ |33111110111118 1卜6 ^ 3 ( 3 561 1010 568016013 1131113出6 0161110(1 0『1^6014^^6 IX!"""。"^.

1461*6 I ^111 ^1301155 1 1 0 ^ 1卜6 10131115 31*0

56160163 &I 6&0\\ 316(3 0『1116 36^1161106 0 ^ 1^00「51乂6 ^311111011 !!!^. I ^111 ^00115 0 0 1\^0纖\\^乂, 0『610317: 5|3|11:3 0『1116 I叩1115, &111101^11 1111-66-^3^ 3 1 1 ( 1 議 出 - 篇 丫 5^11(3 0311 1)6 ^1^0^1X1^6 ^1;06『35 \^611 I I I 3(^1出011, 1(16 0115011551011 3551111168 111311)16

13 1510317^

^

50161^)1-156


122

?1-6(110^6 1^0滅"宕^1(11 8/18 ^6̶36

4 0 脚 3

4.3.1.1

00(10018011(110118^ 81481^1688 ^^1100(10113

141^:

1^0016 1 3 8(3111 。81119 3 8 1 1 1 3 1 7 5(3111 3 6 3 1 - 0 1 1

10 & 5 1 0 3 1 7 5口 11仁1^0 1)1^110^65 6 1 1 1 6 ^ 0 & 0 0 1 03011

^ 1 1 6 0 3 0 10(3111 311011 3 8 11011361101(1 1 0 0 0 0 1 6

15 1136(1 10 13311111011 1116 1^600^5 111(0 1 ^ 0 3『011卩5, 3 3^1111117^ \01116 311011 3 3 1116 11011561101(1 10001116 0『510,000 0 1 3 ^ 1)6 01103611. 丁 116『6001^5∼&卜111001116 1655 111311 1116 8^11^117^ ^?/^ 31*6 5 6 0 1 1 0 1116 16& 6『300|1, 311^ 1116 1*6001*^8 ^1111 10001116

3

0

0

『60)1131 10 1116 8^11111118 ^31116 31*6 5 6 0 1 1 0

1)16 1*12111 ! " I I I & 0111111-^3^ 5口1!1, 0101*6出30 ^ 0 1)1-3001165 0 1 3 ^ 6 0 1 6 ^ 6 『 「 0 0 1 3 0 ^ 00(16, ?0「 6X8011516, 1116 111001116

&616 001||^ 1?6 ( ! ^ ! ^ 1010 1*210865 5110)1 &3 50~510,000, 510,001̶520,000,

520,001̶ 530,000, 6 1 0 . 八 3 016011006(1 63『1^『,『0『1116 13111^)0363 0『31011)1101^ 80(5 。131*1^, I 00011116 ^ 6 (!^^!^^&!! 10 1^/0-^3^

3(31115 301^033 3 0 10^111

10 0^6^ 10 3^111 3 0 ^ 3631116111: 0『5116-56^11601 0『1116 3&I& 361 &1 & 0 0 3 6 , 50161^1^136 IV! !^^!" 03101113(65 1)16 ^01111 0 ^ &11 0311(11^3(6 313111(1113 乂311165『0『63011 0 ^ 1)16 !^!? ^, &11(1 56160(3 1)16 5 ^ ; 1

出6

!11乂0^65 11)1(111^5 \\\& 1)651 5(^(11)13 乂81116『0『03011 0 ^ 1116 3^311&1)16 1111)11(5 311(1

111311631 ^01*111,

111611 001111)31*1113 &11 0 『 ^ 6 11)^11(5 10 5 6 0 ^1110[! 0 0 0 * 5 5|?1& 13 1)651 丁 116 !!^出。^] 0^03101113(1113出6 \^01111 0 ^ 1116 5口III 0 & I I 1)6 36100(6(1 !)乂 1116 056『,丁 11656 1116(110(15 3『6 0115011336(1 I I I 1卜6 ^ 0 1 1 0 ^ 1 0 2 360110115.

7110 1)1*00633 0 ^ 56100(1113 1卜6 1)651 ^1116『0『630(1 111(3111: 13 ^6(61*111 !^^^. ^

00(151515 0 ^ 1 ^ 0 316口5, 1)1 1卜6 ^1^31 316口, 1卜6 1)631 3(3111(1113 1(1 016 560011(1 516口, 1116 6 6 3 1 ! 叩 1 1 1 細 0 1 ^ &11 I 叩 0 1 5 13 8616016(1

001111331-1113 1116 \^01111 0 ^ 1卜6 1)651 5|?111 0『6&0^1

56160(1113 1116 10口01 \^11086 1)651

&1)16

1116 1?631: 3^)11(3 0 ^ 1卜6 0!116『3, &0(1

3^111 ^ 6 1 ( 1 5 出 6 1118卜651讽01111, 丁116 00(16 15 3口 111 &11116 13651 8^1111111^

^!1^ 011 1116 1)651; !11口71113 口 1*00655 0311 1)6 11111511*^6(1 6 ^ 出 6 ^ ! ^ ^ ! ! ! ^ 6\&1111316.

811(313036 11161*6

100 10(31115, ^ [ ? ^ ^ ^ ! ! ^ ^ 6乂乂^ ,义 ,…,义 2

|00

, 丁卜6 11*66 31301*1(11111 5131*15 ^ 1 出 1 ^ 1 1 1

〈 0 , ^1161*6 ^ 15 3 ^ ^ ! ! ^ ! ! ! ^ 乂3106

6X811111165 311 0311011^3(6 3(311(5 0【1(16『0011

50016^1161-6 1 ) 6 1 ^ 6 6 0 出 6 111III 1(11 I I I I I 311(1 1113X111111111 乂31116 1116 16& 01111(1 00(16, 31101 &11『^(^『(^

1 1 ^ 6 《 2

0 8

,八11 !"6001X15 11131 1 1 ^ 6 0 1 0

& ! !^ 01111(1 1106丁 6

9

^ 0 ^ 0 10

116 31^01*1(11111

^ 0 6 5 11^011〇11 &11 0311^(1316 ^ ! ? ! ! ^ ! ! ! ^ 乂311165 0 1 1 化 6 531116 !11)3111 &11^ 5 6 1 6 0 ( 3 出 6 1)631 51)11(11113 1.611116 1)651 5口 1^1(13 06X1 1 0 ^ 1 1 ^

2

0 0 10口01

^ 71115 131*00655 15 ^6^16^

1 ) 6 ^ , 丁 116 81301*1(11111 1*6^63(3 ^ 6 33016 口!"00655 0 0 ^ 1 6 0 0 10131115 ^

2

5

^

9

^

4

5

^.., ^

|

0

30(11卜6 001763|30110|1113 1)651

0

3^11111118 乂811165 (^,(^,(^,…,(^ 3 1 - 6 化 " ! ^ . 9 3 ^ 1 0 ^ ^011001 1116 1)631 5口11扛108 !11 口"I, 1116 31801*1111111 11160 0001(531*63 111656 3^11(3 10『111(1 1卜6

『0『630)1

\^11056 1)631 3^11(11113 乂31|46 8 ^ 6 5 3

8^111 0『1116 6&I& 11130 1)16 1)651 5 ( 3 1 ^ 1 0 8 乂3106 0 ^ 30乂 0也6『1(1 口01. 311(313086 0^ 13 1116 1)651 8口『0『101)111:

31101 3111)15036 ^

|

0

18 0110360 3 5 1116 13631 10口111 11|3011 ^111011 10 1)336 3

5^)111 0 ^ 1116 0036, 00056^116111!)' 1)16 110^6 15 133111(10116(1 1151113 5

^1111X、。 〈 0

1 0

31-6 5601: 10 1116 1 6 ^

比口 111《。^八11 1^600^5

0036, 311(1 &11『6001x15 ^1111 ^

0)111(1 1106&. 71115 131*00653 15 1*6)363(6(1『0『630(1 1\06八出1?6『6111

1

0

^

0^ 81*6 5 6 0 1 1 0 1116

0 1 3 ^ 1)6 01105611 31: 3出疗61~6111

1\06&. 1\0\^ 31*6 1116 3(311(3 (;。!!!!^!^」10 0161611111116 1116 1)651 3|31113? 丁0 805∼6『1^15, !1 !5 0 6 0 6 5 5 3 1 7 1 。 6X31111116 1 1 0 ^

1116&3111*6 1116 ^ 0 1 1 ( 1 0 『 0 \ \ 3^111.


4: 81(11(11)1^ 0^1X10(2 丁"" \40^\^ 10

^1?017^

011(1尺^

123

1^6331^1门9 1匕6化0「1^1 0 ^ 3 3^111 (^^!:!!]!!^ 0116「1011 ? 1 - 0 ) 3 6 ^ 0 1 1^16 06018;0^1 丁!"66 1 ^ 0 ^ 6 10 5 0 1 6 1 ^ 「 1 5 6 1\1106^

4-3.1.1

丁卜6 5151111111^ 11161110^ 115601『0「 |3311;1;011;11& II化^313 !5 ^6161-111111601 1 ^ 1116 ^^1110 0『1116 0 ^ 6 「 ; 0 0 卩「0卩6∼!11 1116 8|5111110〇1^1116 56011011 0「 11化^1~0|3611165 ^31101. (^&!^^呂111 11化0011111111 10 1116 「^!11 0^1)10 01-1(61-1011卩「0|3611^ ^111^5 11(5 3 11161111 0『廿16 乂3「;005 1116(110(15 0『5|?1111111^ 35 5110^0 !0 0 1 5 ^ 1 3 ^ 4.4^ 013^13^ 4 . 4 01':161100 06181111 ? 「 。 ^ ―

。56 1叩山0^06 捕咖訓031690^1031 5126

?10^

1116 131*^61 ^31*13610 15 1)11131^ 0「 (^((^。「[(;" XV 1(11 111012 111311 1\^0 16乂615,【1化「6 3(3(5「0301165厂0!" III635III"!]1^ 1116 \^011:11 0^3 5(31(1:

1\^0

擊 1116 1^「66 0「 563^311011 3 ( ^ 1 6 ^ ^ 1卜6 5卩11【 ^

1116 1111(3111-11^ ! ^ 110(1011 30111^6^ !))' 1(16 1门6016「!5「156 IV! !06[-,【1化^6^1~06 0「 56;53「311011 15 17103511^01 1?^ II化7^-^31116 0『1110 ?031-5011*5 1651, 311^ 1116 1|311「11 乂 1*6(5^011011 15 11163514!"^^ ^!^!^!" 1^^ ^ 1 1 ^ 0 ^ (^(^"(^!!^】0!" ^ 01111 「(:爿"";。!!. -

1110 13!'^61 15 00111111110115^ 6016「!3「156 1^!1|化「 11163511「65 1116 、乂0111, 0『3 5|?111 ^ 311 ?一1651, \\111011 10515 「0「 0 化 1 * 6 6 0 卩 ^ ^ ^ ^ ^ ! ! 0「 1116 (:卜;^ 1101^5, 丁卜匕0|13|316!' 100^ 31 1)1656 111635111-65 0350 ;

^

035^

4 . 3 . 1 ^2,八11/163311^1119 4.3^ 1.2^

1 060!'66

\^01^11 0 1 3 3^1!!: ^ 1 1 6 0 1116 丁 3 「 9 1 18 8 1 0 3 ? 7 6

。, 56(33!'311011

八11^ 1∼0-∼3^ 5^111 ^[!!!^!!^ 3 110(^0 111(0 1\\^0 (;卜!旧1100165. 1(9^!'0/7^ 15 3 111635111^6 0『1,0∼ ^1^61-6111 111650 01111(1 110^65 31*6『「0111 63011 011化!", 3113 &0111 1)16 |53「61】1 110^6 !11 16「1115 0^1116 00111^05111011 0「 1)16 0135565. 丁卜6 ^ " " 出 ^1^6^1106 1^61^2611 1 1 化 ^ 0 01111(1 110(165, 1116 " 化 ^ ^ ! " 1116 ( ^ 「 化 0 「 56^31*31 !011 3 ( ^ 1 ^ 6 ( 1 6 ^ II化5^)111. 311(1出6 !^6「 II化5(5111 15 00II51010^(5 10 1)6^ 5

1

6 3 1

6

丁0 366 |10\^ 108̶0111, 13 03101)13(0(1 31 030)1 ^))! 1(11113 ^116, 001181(^6!' 11113 0X3111^16^ 1^61 11)6 13!"〇61 1)6 "(^(""伫.311(1 161 !1011x0/101^ 1/1001)1^ ^ 311 !11^)111 (^^^ 131101017 ^^! ^!?^). 1^61 62011 「0\乂 0『11化^3(3 561 !"6口!"656111 3 「。(^「廿《0!" 0^561^311011 ^ 「「0111 1)16 ^3^3 561. 丁31)16 4 . 2 5110\\5 3 ^311131 V&、V 0^1116 ^313 501 III XV111011 I 5|10\^ 011∼5616016(1 1-200^5 0001 1)16 ^^1^1 561 311^ 011 !)' !^0 00111111115. I 3 1 5 0 561 11(3 131)10 50 1卜31【1化「。(;。!")^ 3「6 III "〖「^^!^口^ 0「36「 0^111001110 10 1101^ ^1501155 1116 0011011 0「 5^1!11!|1^乜化!'000!'015 011 1卜6 53515 0『111001112. -


124

?1-6^1011^6 1^0^611^

^///? 848 ^16^186她6广?^001\00\

80114(10173 /(^ 81481^1688

^1100(10118

731)16 4.2

丁卜6

6^.1&

1^681)01186

11011861101^ 1 0 0 0 0 1 6

1

1

38,000

2

0

510,000

100

1

512,000

1200

0

814,000

10,000

1

8150,000

3110^11 I I I

7&616 4.2

0311 1)6 5口1!1 &I出任61*601: ^311163 0^11011361101(1 10001116.八1: 630)1

111

5^111(1113 乂31116, & 2 x 2 0001:11126110^ ^ 1 ) 1 6 0311 1)6 0 0 0 3 ^ 1 1 0 1 6 ^ 3 3 5 1 1 0 ^ 0 9

5110^5 311 6乂301口16 0 『 0 0 6 5口

丁116 00111111113 ^6^1^6561111卜6 1 ^ 0

丁31)16 4.3,

110(163

73516

4.3

1*65011 6*001 1卜6

5)3111, 31101 1116 1*0^8『6口!"656011)16 16乂6!5 0? 1卜6 !3^6^

丁3616 4 . 3 10001116^310,000 116515011^61*5

10001116^810,000

"!2

"口

1^00-1*651)011 (!^^

"引 "∼

70131

VI

^2 ∼ 0『56口31*3^00 &0\\16\^6 ^

丁0 385655 1116 ^6^^ 0111-3^1131-6 ( ^ ) 111036

3 5口11仁11 15 0 6 0 6 3 5 3 1 7

313115110 3.1X6 X&511116 0011 11^1)01116313 1^11116

1

0

0^)0111^6 1*16

口『0口01^011 0 ^ 1*651300(161*3 3010118

111001116 1635 11130 510,000 15 0 0 1 311^61*6111 & 0 0 1 111056 ^1111 111001116

10 510,000, 丁卜15 0 3 0 1)6 ^『1忧60 3 5

//^

: ^

二尸 , \ ^ 6 化 ^ 2

, 為 ^ - ^ " , 311(1^ =

1

1)1166^ 1116 111111 11^1501116515 出6 6X1360(6(1 ^ 1 1 1 6 5 ^111 ^ 6 9

10001116^310,000 ^6313011(161-8

0『1(16

111001116^810,000

11130 0「 60)1131


0110^

4: 814脑"名060131017

I。 ?^101 ^68^01186

7^66 1^0(^13

0^飽化125

丁 116 0111-8(11131-6 3(2(13(10 15 03)0111^6(1 3 5 ^6110^5:

1=0 7=1

^!]

丁卜6 ;7^31116 0 『 1 3 『 0 0 0 ( 1 6乂 301^1112 1116 6(^1131100:

尸(义

2

〉 0010141016(1

I 口 " " !!)'!301116515 13 1 「 1 1 6 】 ^ 八 0 ( 1 1^10 10^0^

15 3 1 0 1 0 3 1 0 1 ) 1 3 ( 6 ( 1 3 5

二 ^ ^ ( ! ? - ^ [ ! & ^,

10^0^

丁卜6 1耵86「 1116 10^01-1^2

(&0(1 1116『6『0『6出6 51113116「出6

1-611116 1。^0她081001316(1

& \ 1116

1116 5^)111.

,八001116『5^1:1 13 0 1 & 3 6 3.11116 116X1

1116厂11*51 5(3111 1)6

0^

1000016 ( ^ ) ! " 6X3011316, 512,000)5 3 0 0 0 1 1 0 8 6 0 0 ^ 0 6 ) 6 !5 1 1 1 ^ 6 811(1 1卜6 1 0 ^ 0 1 1 1 1 13 08101113(6(1 I I I 1卜6 ;

53010 XV&^. I 0311

11^

2

, 1『出6『6 31*6 1 0 0 出 5 乜 0 0 乂31065『0「 10001116 ! 0 化 6 ^ 3 ( 3 561, 99

00011026110^ 131)163 0 3 0 1)6 01*6316(1, 311(1 1116 1。^01111 00011311^(1『0『63011. 丁 116 0 3 1 0 1 1 1 ^ ( 1 ^ 1 1 1 6 5 0尸 1116 10^0^

3「6丄)^ ,

,…,丄^ , 丁 116 5口1!11^1 2 6 0 6 1 * 3 1 6 3 出 6 111311631: 1 0 ^ 0 1 1 1 1 !5 5616016(1 9

5 0 | ^ 0 5 6 1116 1)651 5^;1^08 乂31116『0「 1门00016 15 515,000,

3 1 0 ^ 0 1 1 1 1 0『20,5, 1^0\^ 0005;016『1卜6

0 6 X 1 ^31*131)16,八0^ 1『11161-6 3「6 67出51!001 乂311165 0 ^ 八 0 6 !0 1116

1116 013X111111111 1 0 ^ 0 1 1 1 1 15 301601001 & 「 八 0 6 , 311^13056 1116 1)651: 5^111(1113 ^31116『0「

1116 5(3111: 八06

3 6 ^ 66 3^1113 31*6 00051^61*6^;

15 35,

& 1 0 ^ 0 1 1 1 1 0 ^ 1 0 . 2 . 1 『 ^ 6 &1\6 111001110 31*6 1卜6 001乂 111^1115 I I I 1116 6&1& 561, 111611

10001116 13 3616016(1 10 5|?1111116 0 0 ^ 6 66631136 II II&5 1116 1)651 1 0 ^ 0 1 1 1 1 乂 3 1 0 巳 1 0 0出6「 \^01"(!5, 1)16 6 6 5 1 5(3111: 0 0 1000016 ( ? ^ , 。 。 。 ) 1135 3 102^01111 0 『 1 0 . 2 . ?161106 1116

1 0 ^ 0 1 1 1 1 卩 。 ^ ) 11130

361 \^011101 1)6

15651 5|?1!1 0 0 八 0 6 〈35〉

3

3.1 5 1 5,000 0 【 1 0 0 0 0 1 6 . 丁卜15 5^1!1: 0311 1)6 0&\\^6

1116 1)651 0『1116 13681. I卩11161-6 ^ 6 1 - 6 200 !01)1115 111化6 3&13 561, 1116 |31^00635 0 ^ 1111(11113 ^

1 6

1?651 3^111: ^0111^ 1)6 |361^01111001

200 1117165 ^0110&『0『63011 ;!!)? !^) 311(1 ^0111(1 1)6 1*6^63(6(1 3^3111 3.1 63011 110(16 11131: 15『1111|16『 5|?1|^ 83011

\^0\\\6 1)6 6X301111601 1116 1)681 5(3111 5000(1, 30(1 1116 0 0 6 ^1111 1116 11^11651 !0^0/^/1 ?

0110560 3 8 1116 1)651 0『1)16 13681.

丁0 1136 1116 0111~80|113^6 !1161:110(1 0 ^ 011005108 &

仁3611)16

廿 ; 1 ^ 1(1116 01*2(61*1011 ^0^61^

1116 1)60181011 1*1*66 110(16 10 ? ^ ^ ! ] ! ^ . 丁卜!5 13 1卜6 ^6^1111:『0『1)10317 ^ ^ ^

4.3.1.2.^.2

1/71 口"⑩^6^110^1011

5

0『

'

35 3 11/1633111'& 01 000(111653 0 , 3 3 ^

10 ^01617)1*186 \1!116『,II 13 口055出16 10 0110036 1113X111)121113 / 零 " 々 0 ^ ^6(^11011011 2& 3 0 3110111311^6 10 013X101121118 1110 ^ 6 ^ ^ 6 6 0『56口3『31100『0『30160(1113出 5(31115 0 0 3 6

1)651

&1\ !0|3111, 30(1『0『5616011118 ^ 6

^ 1 1 3 1 13 10113111*1^? 101^111*11^ 0 『 3 0 0 ^ 6 18 1)16 36&『66 0^116161*080061^ ^!111 1*65^601 10 1卜6

00111(3031(1011 0【1卜6

16^613 0『1卜6

I&^1

\3^13^\&.

1^ 110(16

V 13 5口111 1010

0\\\\6 110(^63 0 ^6 6, &1\6

1^ 0?0 3 0 3 6 ^ &「6出6 口『0130111005 0『1^600^5 I I I 00(165 0 311^ 6, 11160 1卜6 (16^I'6^56 111 1111(5111*11^ 15 《V〉 一 ^ / ( ^ ) 一 ^ / ( ^ ) , ^1161*6 / ( ^ ) 13 1116 !!!!))""!!^ 111^6\ 0『110(16 V, 30(1 /(^) 311^ /(^) 31*6 1116 111^6X65 0? 1116 0^\\6^ 110(163 0 3001 6, 「65156(^1^61 丫. 1|1 501617)1*186 ^1!06「 1116「6 31*6 1 ^ 0 01635111-65 0 ^ !01|511『11^, 0 0 6 0163501-6 0^116(1 1)16 01171"/?";77^ 7"^&X, &II01 1116 0出6「 03116(1

5"化0口》I

016501*11)6 111636 11163511!'65 ^ ^ ^ .

丁0 5卩1111116 00(^0 V 1010 0^11101 0 0 ^ 6 5 0 3 0 ^ I? 1)350(1 0 0 出 6 ^ ^ ! ) ^ ! ^ 乂3106 0 『 3 0 10^)111 丁「66 31301*1(11111 0X311111163 &11 0311^1(13(6 5(311(3 0 ^ 1110『01111

〈0

30(1 X, ^ 0

,∼116「6 0

, 1卜6 \5 3


126

?1-6(1101^6 1^61^

^///? 8^8 ^16^186她":?1-0^1001801^10)73^

8148111688柳!!0^0173

1*631 ( ! " ! ^ ! ^ ! " 6 6 ^ 6 6 0 1116 III111111111111 ^6 013X10111111 乂31065 X、 〈 0^ 8 0 10 1116 16& 01111(1 110(16 311(1

^

, 711056 1*6001X15 11131

10 1 1 1 6 『 ^ ^ . 3111)1)036 11161*6 31*0 1 0 0 0 3 0 ^ 1 ^ 3 1 0

, 丁 116 030(11(1316 3^1111108 ^311163 9X^0二

3(31101118 乂81065 0 0 出 6 10(5111

1,2,∼,100 , 丁卜6

31801-1111111 0 0 0 1 口&1*65 1111|311^ 1*6(111011011 0 0 111036 1 0 0 5)31115, 3 0 ( 1 36)6015 1(16 0 0 6 11131 301116^63 1卜6 21*6316311-6(11101100 &5 1116 1)031 5^111.

4.3, 7.2.4.2, 1 I卩/?,

01/111爪口"『!"

1^&X

15 1116 ^1*0(50111011 0 ^ 1-681)011(161-5 111 3 00(16, 311(1 /?。 15 1116 口『013011!011 0 ^ 000-1*65口(^(^巧,1116

01111 / 零 " " / 亇 / " ^ 0『11131 00010 13 ( ^ ! ^ 3 5 / ( 尸 ) : 1 一 /?' -

, 1『1^0 1*6001^5 &『6 0^105611 3 1

2

1*311(10111 (∼辻!! !^…&^^!!^!^) & 0 0 1 & 0036,出6 ^01)31)1111^ 111311)0111 31*6 !^?!^!^^!^ 15 ^

, ^ 1 1 6 1116

&1)1111^ 11131 60(11 31*6 !!^^-^^^^!!^^^ 15 /?^ , & 0 3 出 6 (?①!)&!?出矽111311\\6^ 31*6 6!1116『601卜 ! ^ 网 & ! ^ 01- 50(11 000~1^6315011^61 5 15 ^ 十 ^ \

, 961106 \~ ^\ ~ ^\ 0 3 0 1)6 ^ 。 『 ^ ( ! 3 5 1116

1

131*01)31)1111^ 1^13.1 30乂 1\^0 616016015 01103611 &X『80(10111 ( ^ ! ^ 1*6口^仁^!!^!^) 31*0 31伃6『60伫?0『610317 131^615, 1116 0)111 111016乂 51011)11^68 10 2/7, 〔1 一 /?,〉,八911『6 00(16卜35 & 01111 1113X111111111 乂3106 0『1/2 ^ 1 1 6 0

1)0111

0 ^ 26『0, I I 1133 &

0133365 3『6 6^11^11^ 1*6^1-656016^^ 丁0 1156 11113 0『116『!011, 5 6 1 出 6

81)1111103 ^1116 01*1(611011 ^ 1 * 0 ^ 6 1 ^ 10

4.3.1.2.^.2.2

5/1^0^ I

5 0 1 1 - 0 ^ 13 &001|16「 0163511『6 0『1116

^(/?/)『0『

0『& 1\06&. I I 13 ^6^106(1 3 5 /(/?) ^

1=0 510317

八 0 0 ( 1 6 111311138 131^6「 601『0口乂 11130 3001116「 00(^6 15 0101*6 116161*0经606005 3001

1116『6【0『6 1633

丁 116 1-31*1^ 0 『 3 0 6乂601 18 016&8111*0(1 3 3 " ^ ^ ( / ? / ) .

1『30 6乂6111 15『&1*6, 11: 17)03113 1)16 ^1*01)31)111^ 0『

1(3 000110-61106 13 10∼,0^011510|6『1卜6 6乂601 ^68^01186^ 811|3|3036 1116 ^1*01)35111^ 0 ? 1*63^01136 1)1 3 0 0 ^ 6 15 0^005^ 71160 1116『3『!1^ 0『『65口0056 13 一 1 0 8 2 ( 0 . 0 0 5 〉 二 7 . 6 4 4 , 71113 15 3 I 『 6 6 乂 6 0 1 丁卜6 ^1*01)31)1111^ 0『000-1-631)01136 13 000乂6『561乂 0 . 9 9 5 ; 1161106 1116 1^1*1^ 0^000-1*65(501156 15 一 1 0 8 2 ( 0 . 9 9 5 〉 ^ 0 , 0 0 7 2 , 八 0 0 3 6 出 & 1 1 1 3 8 &『65口0056 1*316 0 ^ 0 . 0 0 5 13 1635 11111)111-0 11130 3 110^6 11131 1135 6^1131 口『0口0111005 0『1-651)011 ^61*3 &1\6 !^!!-!^。!^!^. 丁 11115 ~\0^ (|? ) 2

!'&1*1^ 13 111^11 3001 1 0 ^ ^ 6 1 1 1116 1*31*1^ 0『出6

/(/?) ^ ― 【 0 . 0 0 5 、 1。8 ( 。 ' 。 ( ^ ) 十 0 . 9 9 5 2

|

13

\\^60仂6

15 1 0 ^ ^ 丁 116 601『0口乂 0『11115 00(16 13 35卩0110\^5:

1。8

#

2

^ 0.0454,

( ^ ^ ! ^ & ^ ! " 3001116『1106& 111 ^ 1 0 1 1 1116 ^1*01)31)1111^ 0『1*65(301156 ^ |3!'01)31)111^ 0『000-1*631)0036 ^ 0.5, 1116 6 0 1 1 - 0 ^ 0^11113 00(16 15 / ( 尸 ) ^ 一【0.5 # 1。8 ( ( ^ ) 十 0 . 5 # 1。8 ( 。 ^ ) ] 二 1 . 2

2

110016 1 ^ 1 13

131-6(1011111131111^ 0 0 0 - ! " ^ ( ^ ( ! 加 ( ^ ! ! ! ! & 131X^011100 0 『 0 . 9 9 5 〉 1135 3 0 6 0 ( 1 * 0 ^ ^ 1 0 6 0尸0力454,八 00(16∼!出30 6^1131 ^ ! ^ ^ ! ^ ! ^ ! ! 0『1^65|3011^61^

000-1*651300(161*8 1135 6 0 ^ 0 ^ 6^X13.1 10 1 ,八00(16

\\\31 1138 8\\ 1*651)011(161*5 0『&11 !!^^-^? ^^!!^^ 1133 311 61^1*0)3乂 6(^X19.1 10 261-0. 丁 11115, 6 0 1 1 * 0 ^ 1*311365 1 ) 6 1 ^ 6 6 0 0 10 1 , XVIII! 0 111(1103(1113 013x1011101 ^111*11^ 311(1 1 111(1103(1113 11)^x111111111 111115111*11^^ 丁0 1156 1116 ^ 0 ^ 0 ^ 111635111~6, 5 6 1 出 6 8 1 ) 1 1 ^ 1 1 ^ 1 1 1 6 01*1(61*1011 ^31*013611^ 10尼!^!^口^


(^。声"4: 81111 出"8

4.3.1.3

060151071 7^66 1^0^613 10 ! "6(^11^63^01786 017^ ^18^

127

3

1\^633111 1119 1 & 6 \ ^ 0 1 1 1 1 。, 3〇口111 \^116111116 1 1

&

\ 18 0 3 1 6 9 0 1 1 0 3 1 ^ 1 ( 1 1 1^0『6 |

711311 丁 " 。 0 3 1 6 9 。 『 1 6 8 1『1卜6 ( ^

6 1

15 0 & 1 6 8 ^ 1 ∼ ; 出 1 1 \ 0 ^ ^ 0|

1 ^ 0 03(6^01*165 ( ! ^ ^ ^ ! ^ ) , 出 6 ^1*006(1111*65 31*6 1116

03

丁 116 0111-8(11131^6 5(3(15(108 ^ 1 1 1 1)6 03)01113(6(1 & 0 0 1 "乂 !) 0 0 0 1 1 0 2 6 0 0 ^ 1 ^ 1 ) 1 6 5 , ∼ ! ! )

15 1116 1111111661^

0卩01111(1 110^65 1)61113 01*0316(1 0 0 1116 1)3513 0『& 0^113111化口" 311(1产15化6 111111166『0『\2^^^\ \^\^\^ ((:^圪径0『1651 丁 116 ^ ^ 3 | 1 1 6 5 31*6 0^101113(6^ & 0 0 1 1(16 01^-5(11131*6 ^13(1*1611(1011 ^1(11 ^6^1*665 0『 ^1-66(10111 60)1131 10 ( 广 一 1^(6 一 1 〉 ^

丁 116 0!01 \1\66\ 30(1 ^ 1 1 1 1 * 0 ^ 01635111^65 0? 101|311『|1^ 0311 3150 1)6 3^(3116(1 I I I 1卜15 0&8&. 丁 116乂 21*6 5!111 口 1乂 6X1^1X6^6『0『010『6 1(130 1 ^ 0 1 ^ 6 1 3 ^03.1I&0『1卜6

4.3.1.4

1 3 ^ 6 1 乂3『1315|6,

1\1638111 1|19 1 ( 1 6 ^ 0 1 1 1 1 0 , 3 8 ^ 1 1 1 \^11611 1^16 丁3『961 13 0011(111110118 1

1『1116 131^61 15 0 0 0 1 1 0 0 0 0 5 , 3 0 ?~1651 15 1)86(110 01635111*6 1116 ^6^1*06 0『36(33111011 &0\\\^\^6 ^ 5(31 !!. 311(313056 1116 1 3 ^ 6 1 15 1055^6^14617^

9

&

& 0 3 出 6 111(3111 13 117001116 \\3^11\^ 100 ( ! ! ^ ^ ! ! ^ 乂31116孓 &, 1^\0 ^1*011^3 31*6『00116(1〇06

丁 1161*6 31*0 9 9 ^0351616 1 ^ 0 - ^ 3 ^ 3)311(3 0 0 1 0 0 0 0 1 6 . 八 I 630)1 5^5111

21*011^ 1133 10001116 1 ^ 6 1 156|0\^ 1116 5(3111: V&]\16^ 311(1化6 01116『2X0X1^ 1135 1000016 ^63X6^出30

0「

691131 10 1116 5|?1& 乂8|116, 1~16「6 13 311 II 111511*311011 , ( ^ 。 ! ^ ^ ^ ! ^ & 口3『6111 110(^6 \^!111 VI 「6001X15 311(1 & 3^111 ^31116 0^510 5 000^八II『6001X15 ^1111 1000016 1633 11130 ^1111 10001116

1)130 0『6(^1181 10

810,000

510,000

31*6 5 6 0 1 1 0 1)10 1 6 ^ 0 0 3 6 , 311(1 111036 00(16, 3 5 5|10\^0 111 「^111*6

31*6 5 6 0 1 1 0 1110

丁116 1116311 0『出6 131^61 13 0310111^6(1 ^0『630^1

^19114.2 ?&1*601 00(16 X議1^『0『1^01X15

1

1^630 0卩1(16 \31^1 :

111001116

,'

〈510,000

2510.000

,'

,'

0111(1 0 0 ^ 1

011(1(1 00^6 2

1^0011^『。!" !^②!^ ^

!^!!!!!^ 0卩^600^5 ^ 7?

\1030 0『出6 131^1 :

\16311 0「出6 131*^1 :

2

10 0「36『10 03(0111^(6 1116 \^01111 0『11115 5口11仁& 0 1 1 6 - ^ 3 ^ 八 1 ^ 0 乂 八 ! 5 111^1 1561*^00116(1, 3 0 ( 1 1卜6 10^0^

15 111611 0^101113(6(1 & 0 0 1 1)16 ?~163^ 3 5 5110^11 ^ ^ ^ .

4.2,


^41)161': ?1-0011001 ?。/"//。"》丫。" 81181)7653 ^^1100110^18

140(1&11^ ^!1/7 & 4 5 ^16^186

^6^10(^6

丁 116 5 0 0 1 0 ^ 3^1131-65 5 6 ^ 6 6 0 1116 8^011153 ( ^ ! ^ (^(!^力13 8 ^ 6 0 \ ^

88^,, 二 丄 " ' ( 义 一 7 乂, ^1161*6

15 1卜6 01630 0『1卜6 131^61 111 1116 /仏。!^" 00(16, 3 0 ( 1

3? 13 1116 0 1 6 3 0 0 ^ 1116 I & ! ^ 改 ^

丁 116 5 0 0 1 0 ^ 30)1131*65 ^1111111 1116 ^011|55 (。卜!化。。(^^) 13 2 ^ 6 0 5 ^

2 ^^,",/" ^

^

^

(

^ 一又.) , ^ 1 1 6 1 - 6 ^ 13 1116 ^ 1 1 1 6 0『1卜6 1 3 ^ 6 1 ^31)16『0『 2

,。1 7=1

1116/

1

1*6001*01 111出6 /'力011(1(1 1106&.

丁卜6 丁 3 1 1 0 1 0『8^1131-63 !3 8 ^ 6 1 1

^

2∼ ^ ^ ( ^ , - ^ )2,

88福^

1116 ?~813113110 15 00111^11(6(1 3 5

1=1 7=1 、

2

-

0

丁 1115 8(3(13(10 1138 311 ?-^!?!! ^!!!^!! ^1111 1 &1\6 0 ̶ 2 ^ 6 ^ 6 6 8 -

、 3 5 、 』 " 0 『 6 * 6 6 ( 1 0 1 1 1 如 ^ !116 111111 11^0116513^

丁卜6 111111 11^1)01116515 5 3 ^ 5 1^3.11卜6!"6 15 0 0 61^16^06

111 1116 131*861 016011 1 ) 6 ^ 6 6 0 1116 01111(1 。 。 ( ^ ^

丁 116 / ? - 乂 & ! ^ 15 8 ^ 6 1 1 6丫尸(尸〉00101*101&6 ? | 0011 11)^01116313 13 (!!!^) : ^ ^ 3 1 1 1 6 ^

八5 1)6^)1*6, 1116 1 0 ^ 0 1 1 1 1 13 03101113(6^『1*0111 1116 ^ 7 ^ 1 1 1 6 0 ^ 1116 ?-1651 3 5 ~ ^ ^ ( / ^ " ^ ^ ) ^ 丁卜6 13『8 『出 1 0 ^ 0 1 1 1 1 6

6

1\6 1^616(6^6

1116 51113116^ ^ 6 / ? - ^ & ^ ^ ) , 1(16

0 0 6 - ^ 2 1 ^ 311&1^513 15 1301^00116(1 & I 63011 ^^!!^!!!^ 781110 0 0 & ^\\61\

1116 5;31!1

311011 3 5 !11001116,出6

108^01*111 0『630)1 5)3111 13 0 ^ 0 1 1 ( ^ ( ! , 311(1 1116 3^111: "!出1116 11&11651 10^01*111 !5 36160(6(1『0『1116 !0|3111

1000016 13 3(^111: 1010 0101*6

1 ^ 0 ^ ! ! ^ ^ , 11160 0 0 6 - ^ 3 ^ 311&1^313 13 ^0116

1 ^ 0 ^ ^ ! ) ? . ) 八 5 ^1111 1116 0111~8^11^6 11101110^ &1\6 & I I 出 6 3(3131163 10 63011 1门口0仁3X1^ 1116 1)631

0101*6 11130

0161110(15, 1扒6 531116 31181^313 15

1)651: 3^111;『0『63011 !0|5111 13 &11001. ?『001 1116 1)651 5口1讧5 0 ^ 63011 !0)311、

& 0 0 6 " 池 出 6 111^11631: ^ ^ ^ ^ )

13『00001, 311^出31 ^111 1)6 1116 1)631: 0『1)16 1)651,

1)131 1 3 出 6 5 6 5 1 5|?1!1: 0 0 1116 1)651 9

丁0 1156 1(16 ?~1651: 5110^11 63『116『III 11115 56011011, 5611116 ?^!)?, ^111011 13 1(16 ^ 6 & I I 1 1

4.3.2

^1116 0116『100 口『0口611^ 10

『0『!0161^31 1 3 1 ^ ^ 6

^ ^ 3 1 1 1 6 ^111811116111 0 ^ 1 0 1 1 3

1/31113 1116 ^6^31111 56^1'!185, 50161731*156

11363 3(^11516(11?~\^\\165 1(1 00!11|33「^3 5|?1辻5 0 0 1116

531116 10^)111 311(1 3^1113 00出斤6『601 111|3111^ 丁∼0 1^(365 0『&01)115(111611(3 &1*6 1113(16 10 ^~\3\\1&5: 8 0 0 ^ 6 1 X 0 0 1 3(1)11511116111: 30(1 1)6^

"̶(!!^!^.


0^100^^

4: 8"脑"宕0601810)1

8011,6『『0。1 一1131!116111

4.3.2.1

丁"66 1^0(^615 10

^1011^68^01756 011(1 ^18^ 129

9X0^X1^

∼ 1 1 6 0 乂00 &1*6 0001|53『!08 1卜6 3(311(3 & 0 0 1 31汗6『601 10)31115,尸-乂31065 31*6 30^113(6(1 10 1 3 ^ 6 111(0 0111131: 0 0 1 &11 1111311(5 1 1 ^ 6 1)16 53010 !!"!!!!^!" 0^16乂615, 10 ^61161^1 5 0 0 1 6 !0|31115 31*6

30001101:出6

7

50016 31*6 01X^(131, 8 0 0 1 6 31*6 !1011^11&I, 3(1(1 01116『5 31*6 1I1(6IV^I-5^^I^(1. 5110)1 3 3 廳 " 。 " ( ^ " 1)14)^6?' 01(63 0 0 出 6 1)61112 〃0《0 1*1011 0油"1)14^61:

0『101*0,1 1)61)13議"0^6^扒甲"&1\6 0

1 1 1 1 3 出 6 聽 ! I 0^6^ !)寧"^^/^/^.

I ^111

?01^ 6 乂 訓 口 ! 6 , 3

?0『1)113 ^ ^ ! ^ ! ) ^ , 0 0 1 乂

0 0 6 5|51!1 13 6^3111316(1, 0 0 ! ) ' 0 0 6 00011036110^ ^ 1 ) 1 6 15 00051(161*6(1, &1\6 0 0 1 乂 0 0 6 1631 13 |561^011116丄 八II !0(30111^6 "聽1)67 11 !^^!". I XVIII

!" 1^1& ^081 12 1710111^18 0311

0^^611^1(6110165

1)115出6 ^61/^146110^

&30^

^ 〃 ' " / 已 丁 卜 1 5 15 3 0 0^11131 \^\3^\&.

& 0 0 1 0 10 1 0 0「

8叩|3056 !1

乂311165 & 0 0 1 0 10 10. 60161^1^156 \4106「 00011331*63出6 5口 1!1 乂311165 0.5, 1 ,5, 2,5,∼,9.5^ 丁60 0 0 0 1 1 0 8 6 0 0 ^ 121)165 3『6 0 0 0 5 1 『 0 0 ^ , &1\6 116006 160 0111~8^113^6 ^ 1 1 1 6 5 31*6 ( ^ ( ; " ^ 。 . I I I 0出6『 \^0^5

5

160 16515 31*6 |361^0011601 0 0 11115 10^)01『0「 5 6 1 6 ( ^ 0 8 1^\6 1)631; 5(3111,

811^13056 1)16 1^ 3^111 0 0 1116 ^611001116110^ 乂3「53616 1135 3 ;7^^1116 ^ 0^ , ^111011 0 1 6 3 0 5 1\\31 ?

0(2101410(6^^

& 0111-3^11&!"6

11)^)01116315 13 1!"116〉 ^ ^ ^ I 。 01116「 ^ 0 1 ^ 5 , 1116 ^0531)111^ 0^巧!!出!^

I

V 11130 1卜6 0310111^16(1 ^

3.1 1*30(1001 15《110(16!" 1*16 0 1 ) 1 1 卜 丫 口 。 ^ ^ . 丁卜6

^51-01)31)1111^ 1)131, 0111 0 ^ 1116 160 0111-8(11131-6 16515 011 1116 ( ! ^ ! ! ! ! ^ ^ ^ ! ! ^ \ ^ \ ^ \ 3 1 16351 0 0 6 0^1116 16515 ^161(15 3 & 1 5 6 口05!11乂6 36(^51011 ( ^ ^ ^ ^ ^ 6 1*6)6011116 1111)1 !!)'!501116515 ^11611 & 15 30(11^11^ ^!^!^) 10 15 1 - ] " ^ ( ! - ^ ) ^ 丁卜13 & 0 1 1 1 ^ - ^ ^ 5 6 6汀0『1*316 15 0111011 1 3 「 经 6 『 1 1 1 6 !11(11^1(11131 61^0『1*316 0

^!

-

6X3111(316, I犷1116 !11(11^(11131 6仃0『1*3(6

13 1 - 0 . 9 5

1 0

^ 0 . 4 0 1 . 丁卜!5 11163115 11131:

^03511)16 ^ ) ) ! ! ! ) , 11161?~\/&\\165

^ 0 0 63011 1651 : 0.05, 1)160 1卜6 ^30111^-^150 6汀0『1*316

160 乂011 113^6 111111(1^)16

00011)31*150115 ^01\6『0『63011

11110101*63(11113(6 1116 1*15^ 0卩1^6|601102 ^ 6 01111 11^01116515 ^ 1 1 6 0 !1 15 ^ 1 1 1 &. 丁 ! ^ 化 & 巧 ,

01631*1^ 1116 0101-6 1305511)16 5(311(5 & ^31*131)16 1135, 1(16 1655 300111*3(6 1116

出 6 6 6 5 1 5(3111 0 0 1 卜 6 編 忖 0 ^ 6 ^

/ / " 《 " 《 " ( ^ ^0^10612

^ 1 1 6 0 ^ ! ! ! ^ ! ? ! ^ ^ 6 1)651 8^111; 0 0 1)16

1)1^6^ ^0^10,1)16^ 1116 10^01"(!18 1166(1 10 1)6 3(^113(6(1『0『1卜6 V & ! " 1 0

^ ^

1116 0 3 5 6 0 『 1 ^ 1 6 賺 " 0 ^ 6 ? '

0 ^ 5|?1!15, 0『16515, 0 0 63011

1)1^6^ ^0^101)111161-6 13 0 0 1乂 0 0 6 1631 &0(1 116006 0 0 9

301)113(1116111 13 0 6 0 6 5 5 3 1 7 ^ 8111: 1)1 1116 0 3 5 6 0『1卜6 ^611^146110^

V(!"!。^, 1116 1)651 3^111 13 0110560 &0111

& 861 0 『 1 6 0 5!?1&5, ! ^ ^ ^ & ^ , 1 0 8 1 0 ( 1 0 〉 15 5116^301601 & 0 0 1 1116 !0^0"出0『1116

^611001116110^

乂3『!31?16,5 1)651: 5(3111. 10 8 6 0 6 1 ^ 1 , 1 『 & 门 1 1 3 5 1*1 5卩1!15,出60 ^ ^ ( / ? ? ) 15 5111511-3016(1 & 0 0 1 1)16 10^01-1)1 0 ^ 63011 5口1;1 0 0 1\\2^\ \^\\\^ 71113 ^115(1116111: !3 03116(1 1116 8017/6^0111

丁0 &口口1乂 1116 801^/6^0111 ^ ] ^ ^ ^ ,

^145^6^1^

|5『0|3611^ 10 \ ^ 8 , \^111011 15

5611116 0011(61*1*0111

1116 ( ! ^ ^ ! ! .

4 . 3 . 2 . 2 71016 。, 1(338 ― I卩乂011 3611116

11811116111

9X0^^

0 ^ 1卜6 711116 0 『 1 ^ 3 8 8 八 ^ ^ 0 ^ ^

3^11(5 31*6 0011113^6(1 XV1111001: 8011^^0111 ^^81!116111 ^ ^ 3 1 1 1 6 5 0『1116 5^11(5 &「6 0001^31*6(1

4.3.2.3

\0八&6『,111611 1 卜 6 0 『 1 1 1 6

\^ 乂011 3611116 ^1*0^611^ 10 86『01^, 1 1 1 6 0 出 6

8011/^0^1

^115^6^

0 6 口1(1 ^11811116111

10 50161731*156 \1106『,1^^011 5611卜6 8(1111;

!] 118(1116111

0^ 1卜6 1)6018100 丁 1*66 110^6 10 \ ^ 8 ,

11160 ^ ^ 3 1 1 1 6 3 31*6 &^115(6^『0『1116 口"!!!!^!" 0『3006510!" 3^)1215. 丫 0 0 0311 0&11 1卜6 &01)11511116111『0『1116 01101136「 0『3006510「 5|51115 1116 ^6^11 0(1)1181^6111 ^ 1)6031136 1116 33」05101601 ^6(3611013 0 0 1116 ^6(3111 0『 1:116 11*66 3 1 ^111011 1116 5(5111 13

0 6 口 出 1 3 01635111-6(1 3 5 \\\^

0 ^ 151^3?101165 111 1卜6 ^3(11 1X0111


?1~6^10^6 ^40(^611^ ^///? 8^48 &116,156

': ?^00(100180114(10178/。卩81481)1688 ^^1100(10178

1116 0111X6111; 00(16, ^1161*6 1116 3^11(11113 13

10 1)16 1*001 0 0 ^ 6 ^ 丁 116 021101113(6(1 ^ 3 | 0 6 13

010111(5116(1 \^ & ^6^(11 010111口116『,1)386(1 0 0 1116 ^0(3(11 10 1116 11*66 0 ^ 1116 01117601 00(16, 10 & 1 X ^ 6 3.11110 ^613(11-^115(6(1 ^ 7 ^ 1 1 1 6 0 ^ 出 6 5口1!1 ?0『6X301 口 16, 1 ^ 113 3 5 5 0 0 1 6 1113115『10『10 1116 0111X601110(16 11161*6 ^61*6《011!" 5^)11(3 ^0\1^ 8^\\\5 ^61*6 1-6(11111*6(1 10

『『001 1116『001 110^6 10 1116 0111X601:

3 0 ( 1 1卜31

630)1 5^)1& 1 0 ^ 0 1 ^ 6 ( 1 1 ^ 0 1)1*81101165 (^!!!^ 510317 5口1I^5). 10 1)113 0356, 1卜6 36|31|1 11111111|3||6『15

2x2x2x2^ 10 86061*31 1116 6/6^細I"!)"^"『0『61031^ 0110-601:

1106&. 丁 116 0^10111^1601

^1101*6 (! 13 1116 ^6(3(11 0『1卜6

5|?1115 15

^

& 15 &6^5X^6

0111111 口 1161\ 71113

011|11!|5!)'108 ^ ^ 6

11163113出81 &X & 鄉 出 0 『 4 , 1^ 1卜6 03101113(6(1 ^ 3 1 1 1 6 13 0.04, 1116 ( ^ ― - ― 0.04

11516(1 ^ 3 1 0 6 \^01|1^ 1)6

X 1 6 ^ 0.64^ ^111101111116 ^6^(11 34)1151016111, 1116 3(3111: ^0111(1 1 1 ^ 6 1)660 00051(161*63 3^(13110311^

3131111102111^ 8111 3 & 6 『 出 6 3^11511116111,出6 3口111 13 001 513115110&11^ 丁116

&

&4)051111601 0311 &150 1)6 101617)1^16(1 3 5 出 … 出 ! ^ 1116 1(11*63110101 ;7^3|116 ^ 1卜6

!^^!!!!)!!^. 1『1116 1111-631101(1 ;7^31116 313601^16(1 I?乂 1卜6 81^111(1031106

3^113(6(1101?6 0.05/16 ^ 0.003125.八0乂 5口111 " ! 化 & & 361)61*31, 1『01. 1 3 也 6 5;径『103006

1^61

3(3601116(1 6 乂 化 6 81^0111031100 画!"口1^

\^111011 1138 3 ^7^31116 &1)0^6 0 I

0『出6 !0^0"出^

10 11101*6356 1^16 1111*651101(1

1.6^61

^1-0^611^ 13 0.05, 11160 11 18

1)6 !^)^^. 10

31x^6 0.003125

1.6^61

^1^0|3611^ 111611 3 0 ^ 5|31 & ?

15『6』6016呔7^16 6 ^ 6 0 1 0『1116 016(3(11 3(^11511116111 13 10径

〉 二 ^ 1 0 ^ , 0 (?) , 16006, 1116 366136『乂00

10

^ 0 111 & 11*66, 1116 511^016『1^16 5(311(131x1 1)60011103『0『3006口^11经& 5)3111 33 5!@111110311^ 丁卜15 163(13 10 1116 1^6|6011011 0『0101*6

3^11(3

1卜30

^0111(1

1)6611 1-6)6016(1 ^11110011116 66^1\\ &^1』^I5(I116I1^. !"16006, 1116

36口出3(1)113(1116111: 0 1 3 ^ &150『6511忖III『6∼6『!33(11110115, 11161*61^ 111111(1113 1116 5126 0 ^ 1卜6 ^ & .

001111*0111119

4.3.3

3(0^1)1113 1*11165 31*6

1*1*66 6 ^ 0 ^ 1 1 : 8101)1)1119 1111165

8^116(1 ^111*1113出 ( ^ ! ^ ! ^ 1 ^ 8 3 6 0 ^ 11*66 ^6^610131116111: 10 660\6& ^ 6 0 6

11)6

「6001*5^6 1)3111110111118 1135 ^006『3『000118)1 ^ 丁卜15 8601100 ( ^ 化 ! ^ ? ^ ^ ^ 泡 ! ^ ! ^ ^ 3 ^ 5 0『0011(1*0111113 1116 ^ " ^ ^ 0『1116 11*66 1

1116 1^!|11118 1 ^ 3 5 6 0 ^

0

4.3.3.1

0011^0111|19

^6^610^)1116111; 1(1 5)1161^1*156 ^ 4 1 0 6 ^

丁『66 0 ^ 0 ^ 1 1 1 1 1 『 0 1 1 9 1 1 ^ 6 1 ^ 6 8 ^ 1 0 1 3 3 1 ^ 0 3 1 1 0 6

1^61

?!'0|36117 丫011 0 3 0 00011-01 1116 1111(181 512& 0^ 1116 1『66 6 ^ 561(1113 ^ 1116 1111*651101(1尸々3106

1 6

I I I 1*651101^ ^~\3\^

1|1 00161751*136 1^1106『

^81^111^1001106 ^^!) !5 5^601116(1 \^ 56^1113 1)16 81^111(1031106

1^6^61

口『0口611^

0 1 1 1 6 1)60181011 71*66 0 0 ^ 6 1 0 出 6 016311*6(1 ^ ? ^ . ?『001 11)18 ^~\^\\\^ 50161^)1-136 1^106!" 03101113(63 1116 1111*651101(1

10^0^

1111*631101(1 !0^01^

?0!" 6 X 3 0 1 ! ) ! ^ , 1^^011 561: 1116 81^11111001106

0『015611^ 10 0.05, ^ 6 1 1 1卜6

15 8 ^ 6 0 ! ? ) ' - ^ ^ ^ ^ ) , 0『1 30^ 1 ^ 3.1 3 0 ^ 00(16, 00116 0『1116 1(1^11(3 1133 &

10^01"出0『60)1181

10 1116 ^!^?!^化,111611 1卜6 00(^6 13 0 0 1 ^3(11(10116(1『011|161\ 8 ^ 11101*6356 1(16 ^681*66 10 ^111011 1116 1 ^ 0 0\\\\6 110(^63 011151

^601*6381113 1116 1111*63110101 ^~\3\\1^011 61^^

1^6^61

111 0 ^ 6 ! ' 10 0005!016『出6 2 ^ 6 0 5(31)1: 10 1)6 31@111^10311^ 丁 1105, 11*66 ^;『0\\^11 0311 1)6 0011(1*0116(1 ^

4 . 3 . 3 . 2 00111『0111|19 丁『66 6『0\1\^11 III!'0119111(16 5(3111 ^^113(1116111

9^0^0^

八5 171611(10116^ 111出6 560(1011 00》々31116 3^11511116111 0口110115, ^ 乂011 5611116

118(1116111

131*0(3611^ 0『1116 1)60181011 "!1*66 110^6 10 丫68, 111611尸々31065 31*6 ^113(6(1『0『1116 11歸1?6『0『3006510『 5口

I I I ^ 3 1 1 1 0 1 1 1 ^ 1^ 01 1 5 出 6 51^111^1031106 16乂61 5(7601116(1 6 ^ 1116 51^011!081106

111611 30乂 5|?111 ^111011 1133 &

^

0

\

^

1^6^61

口 1*0口611^,

0 I ^?/?//?瞧I"口"6圹13 ^ ^ ^ ( ! . !"16006, 1卜6 6&6^

乂011

8 0 I I I 3 1『66, 1116 5仕1016『1卜6 5(3I1^1^I'^1 1)6000165『0『3006(3(1113 & 3^)1 !1 38 5180|110&01. 丁卜15 163(13 10 1116


0/70/7^ 4: 8""出。8

0601310^1 7^66 1^0(^618 10

6(^101 ^68^01186

1-6160(1011 0『0101*6 5 卩 1 ) 1 3 1 1 ^11110111出6 3(^1151111601, 1*63111(1113 111

0^

/?/^ 131

V |3311出005# 50, 3^11011 !^^

0 0 1)16 81)111八|1』11511116111 |?『0|5611^ 13 3001|16『∼3乂 0 ^ 00011*01110运11*06呂『0^11,

4.3.3.3

0011(1-0111119 1"!'66 6 『 0 ^ ^ 1 III!'0119^1他6 1.63^ 3 : 2 6

丫011 0 & I I 00011-01 1116 ^ 0 \ ^ | 1 0『1116 ^ 6 6 !)乂 56(1(113 I 卜

9X0^^

5126 |3「013611^ 1『乂00 5 6 1 出 6 ^31116 0『

6

1116 |31^0|5611^ 10 3 00011)6!" 311011 3 5 1 0 0 , 11 016305 1)131: 1『3 3^111: 1*65111(8 1)1 1)16 01-63(1011 0『& 163^ ^11)1 『6\^6『11130 100『6001X15, 11160 X[[&X 5)3111 3)10111(1 0011)6 ^^011^\^6^

1^61106, 1卜6 81*0^11 5(0)35 311116

0111X601 110^6^

4.3.3.4

0 0 1 1 ^ 0 1 1 | 1 1 9 丁『66 6 1 - 0 ^ 1 1 111 !'011911 1^6 3(31:1 3 : 2 6 9『。口611乂

I尹1116 ^ 1 0 6 0『1116 81)111 8126 口『0|5611^ 13 5 6 1 1 。 3 II11111 ^ ^ ! 511011 3 5 3 0 0 , 11 11163115 1

\^3 110(16 1138

『 6 ∼ 6 『 3 0 0 1*6001^5,出60 I I 5110111(1 0 0 1 1 ) 6 (^)!^^^!^^]『0『3^1 ! ^ ! ^ . 丁卜6 ^6^31111: 乂31116 0 ^ 11115 0『0!3611^ 15 1^106 1116 \&3^ 5126 " 卜 ^ 卜 1 3 3(3601^16^1 ^ 9

4.3.3.5

1\\6

^ 8126 ^1^0|3611^^

0 0 1 1 ^ 0 1 1 1 1 1 9 丁『66 6『0埘化16『011911 1116 1^3X111111111 06口111

9 ^ 0 ^ ^

丁卜6 1156『0311 &150 0110056 & \ / & 『 0 『 1 1 1 6 1^3X111111111 1)61)111 口「0|3611^ 丁)^5 ^616111111165 1^16 1X13X111111111 II111X11?6『0『经

61161

" "。!^ 0 『 7 1 1 6 3

「001 110^6 !5 8 6 1 6 ^ 1 0 0 0, 30(1 1^16 ^ 11^1*611 0 ^ 1116

1*001 00(16 3!"6 1116 6『51 @6!16『31|011, 610. 丁 116 ^ ^ ! !

56(11(13 『11113 口 ! " 。 ) ^ ! ^ 15 6, 丁 116 1^8X111111111 0

1)61)111 131-0(3611^ 0311 1)6 361; 1)61^0611 1 31101 100^

4,3.4

。1*11111119:

5616011119协6

1^19111-812611

9 3 乂 ! 0 8 2^0^/11 1\\6 1 3 ^ 6 3 1 1303311)16 11*66 (/^^顧I

1131119 ^ 1 8 | | 3 1 1 0 1 1

08^3

"!!^^! 1)16 ( : 。 ! ^ ! ! ^ ; ! ! ^ 0『510口口1113 III165 111 1

1116 1?^10113 56011011, 乂00 0 0 ^ 116601 10 ^114176 \110 1116产/^?/ 5126. 丁 116 131*11111113 13『00655 1)1*0066013 3 3 1(16 1^10X1賺11∼6, 3001 611(111113(6 0 0 6 5^111 31 630(1 &化卩,I? 1116 013X10131 0*66 1135

^0110^5^ 31311 1416^63

311(1 1^1『6010乂6 006 5|?111 31 & 8 ^ 6 0 ^0111^ I

1*6010^611 &I & (! 1^61*601130101 I ^111 ^1 &1\01^ 00111136『0^506-11*065 0 ^ 3 1 2 6 ^ - ! . ^ - !, 1^-2

乂, ^

3 5111?-11^6 0『5126 ^[-1.

1『006 3|31|1: 15

, 丁1105,出6『6 0811 1)6 &

501)41*66∼!出0『8126

?『001 111656, ^ ! ! ^ ! ^ ! ' ! ? ^ 1^1106『3610015 1116 5 6 5 1 3013-11*66 0 ^ 3 1 2 6

1-61X10^108 1 ^ 0 5口1!15 1X001 1116 1113X111131 ^ 6 0 I

X 56乂61^1 5116-11*665 0『3126

^ ?『0111出656, 60161^1-156 1^1106『36160(3 1卜6 6681 511卜1『66 0^ 5126 ^[―1.

1111(11 11161-6 13 & 11*66 ^11)1 0 0 1乂 0116 163^八11^16

0『11115 1?『00655, 11161*6

丁卜15 口『00655 0001101165 1)6 & 36^116006 0『11*665

0『31263 ^ 仏 1 , ^ ~ 2 , ^ ~ 3 , ∼ 1 , 015口13乂 4.7 5110\^5 & 口101 0『出6八\^『3呂6 ?『0化355001316(1 1116 36^061106 0『1116 6631: 0 ^ 1116 ^ ^ ! ? - ^ ^ ? 0『1116 531116 5126. 丁 1105, 1116 56(^1161100 15 3 0 0(3110111111 36^1161106 ?『0111 1卜6 0(3(111111111 56(106006, 1116 1)631: 1 ^ 6 13 3616016(1 0111110 ^ 8 1 3 0 ^ 3 ^ 6 ^ 1

30001-30^ 0『,3116『0311乂61^, 1卜6 1 1 夬 ( ! " " ! ^ 5 0 0 1 6 0 ^ 1)1636 0161110(13 1)1*00653 0『36^0(1113 1卜6 0^111111111 13 II 1113(116(1 111 8601100 4 . 3 . 4 ?1^111-6 4.3 5卜0\^5 1 1 0 ^ 511|5~^665 31*6 01*63(6(1 ^

6

1(1 1116

& 00001^616 6乂301口16,

^111111113 3 131~@6 11*66 3 1 ^11^61-6111 口0!015,

0『0^1仁0『 丁卜6


132

?1-6^101^6 ^40(1611^

&45 ^161-^186應6":

?^001\00\ 8。11^0?75 ^

814311^688

^11001X0118

闩911『6 4 . 3

1x10(16 1

1^10(16 3

1^0(16 2

?1.

尸3

^ 6 7

1^0(16 6

1^0(16 4

尸4,

? 1^111*6 4.3 3110^3 3 11^^0111611031 I I 1135 111116 ^ ^ ^ . 1(1 1116 1113X111131 1^66 11161^6 &『6 163『 110(165 ((^(^^ 4, 5, 7, 8, 30(1 9 夂 8 ^ 1*6010^1118 & 口 111 &X 1116 户4, I ^ 006 1『66 ^1111『00『16^65 5

^ 110(563 4, 5, 6, 3X16 !). 8^『6010乂108 1^16 5|?111 31尸2, I

『00『16^63

& ^1*66

110(163 2, 8, 9,

311(5 7 35 1(5 16^68^ 丁 卜 1 1 5 , & ! ' 6 1^0 5115-11603 0『5!26 4^ 8乂 1-6010^1112 1卜6 3^1118 3.1尸3 30(1尸4, I & 11*66 ^1111 1111*66 1 6 ^ 0 3 ((^!^^^!!径0『00(165, 4, 5, 311^ 3 &3 1(3 163^65^ 8乂 1-61110^1118 1卜6 3(311(3 &X ?1 &0(1尸4, I @61: &001116『11*66 ^11:111卜『66 16^65 0011315(1113 0『00(165 2, 6, &11(1 7 38 1(3 16^63^ 丁卜05, 11161*6 &1^6 1^0 ^^^-!^^^ 0^3126 3^ 8 ^ 1*61110^102出 5口1115 &X ?!,尸3, 2^尸4, I 6

0 0 6 1『66 ^1111 1^0 1 6 ^ 6 5 0005151108 0 ^ 0 0 ^ 6 3 2 311(1

3 &5 1(8 1 6 ^ 6 5 ^

8 ^ 1*61110^1112 1116 5|?1115

尸1, ?2,尸3, &1\6 7*4, I

006 1X66∼&卜001 乂 006 163『,^1111 00(16 1 33 1(5

711115 1116 50^1161106 001131313 0^ 0116 11*66 ^1111 16^65, 1^0 11*665 &11!" 163乂65, 1^0 ^ 6 0 5 ∼ 1 出 1111-66 163\^5, 006 1^66 1^0 163乂65, 30(1 006 11*66 3 3111316 16&^. 81006

11161-0 1^0 511|3~^665 0『3126 4, 0111 乂 006 13 3616016(1 ^ 8101113^ 0116 8116~^66 0^ 3126 3 13 9

5616016呔丁卜15 3616011011 15 1)336(1 0 0 ^ 6 11361'-3111313116(1 0『;16「1011 ^13011336(1 6 6 1 0 ^ ^ 丁116『1031 500|1161106 0011313(3 0 『 0 0 6 ?!!^-!^^ 0『630)1 512&. 丁II !5 56(^06006, &10113 ^ 1 1 1 1 也 6 ^0^11 0『

0出6『35565501601111635111*6 113601 10 0110056 1卜6 1)631 5013-1!"66 0『63011 3126 13 ( ! ^ ^ ^ ^ ^ 10 & 011311 I 。 1卜6 ^6811118 ^ 1 0 ( 1 0 ^ 0『1116 060131011 丁「6巳八5 016011011601 15「6「01151 乂, 015口1&乂 4.7 5 1 1 0 ^ 3 3 0

6X301^16 0『511011 & 011311. 丁0 0)10036 1116 ^1113.1 010*^1, 5016113^136 1^1106『3616013 1116 1)631: 11*06 1x001 1116 56^116006^


〔7";/;&广4: 81111^11^ 0^0IX1011 丁"化1^0(^15 10 ?!~6^101 ^65^0!1^ "'"!

133

\\^1131 15 1110 (;!^化^。" 1156(1 「0「 II呂1)16 1^251 1^2『「0111 3 1111 !!!^!' 0『11~065 0卩II化531116 5120 31 63011 516(3 0『II化56^1101106 ^0501*1136^ 3|30乂6? 0116 1505511)10 01~116!~1011 !5【0 00111 (?^!^ 1116 ^0^\1 0『!116 5115-11*665 III 63011 516(3 01*1)1111(11115出6 56^1121100~ 1110 !'0^1 3550013(6^∼11:11 63011 1^00 !5 0310111316^ 1151112乜化"〃//?/"//.!:"' " ! " 八116「1131^6 01-1161-13『0「 ? 引 。 " " ) ^ 1110 1)051: 1110111(^0 0051 ,11;11 "";2311011, 1111111111 !2311011 0「 1111503^111311011 「3(6, 111111||1112^11011 0卩!'34*6 5^113^^ 6「「0「, 0「 1113\!1111231:1011 0「11||1 1116 0356 0^ 3 00111111110115化1^61, 011,,1|11|231;011 0!"3^61*3^6 50)1130X1 13「「0「 !5 1156^『0「 561601!11@ 1110 13051 1)1 「口「156 ?^!化 | 「, 561 1110 1^1^1110(1 。「。(^!^乂 8111)1100 111 ^16 口「0口61^ ^31161 0〔出0 1)00151011 丁 1*66 110(^0 10 ^ 8 8 6 8 8 0 1 0 1 1 ^ 1116厶5565501601 1\1011511IX |3「0^^ 10 060151011.八、'4;「:1^ 8^1131-0 [「『(^. !\1180|38811|0311011^ 0「 711656 0|5110115 3「6 31251:「1(36(1 360(1011 1「 X 561 1116 、3106 0「 1116八55(^5 016 0 1 ^638111*6卩1*0口61*1^ 10 06。18羞00, 0 6 0 1 8 1 0 0 ^ ^ ! 妙 & 111115: 1)6 61116!'6^ 31 1112 111116 丫011 01*6316 1 1 化 0 ^ ( 3 80111*00 01 5 ^ 561601111^ 1卜6 1111)111 !):"11 110(^6 311^ 0 0

1

0110^111^ 011 口

1003(0^ 10 1110 11^1 0「 1116 000^10118 ^1*0^611^ 1)1 1116 ^ " , 3「0叩0「 1116

|3!'0(36^ 口31化卜^11611 乂011 0|101\ 0 1 1 ^ , 1116 1)00151011 I 1*000881112 ";!!)^)∼0(36115 35 5110^0 10 5

015^13^ 4 丄 0:3|3|37 4.5

工8!"9613 1 「 ?「03 (31 (11111650601510115 | 0601510111/^6191115 | 吣 圖 "邛 ?431116 : 0

^1635111611181111.6^81 :6111317 131^6116^1 01X161-: 6^611116^81 :

1

「0171131 :

0^

08[!061

111611 561601 1110 1)60181011 ^6^11(5 13^ 10 6II16!' II化|5「0^1 11131!'IX^ 011 1116 060151011 X V 8 13匕 611161 1116 「「0厂11 11131!'1\ 311^1 561601 1\13\1111120 111^「出6 13^1 801001 3 60181011『!!!!"^!!. -


134

?1^1\\0\\\^ ^0^611)1^ ^'///? 84^ ^11^1)1-1^編16「

013(3137

4

?!'6^11(^1 80(1(110)1^ ^01~

11^^

^1)110011011^

6

爿 78^9615 I ?1-101- ?1-0^1813111 (163 | 0601510115 0601310^1^6191113 561601 9 ^601510!11(^10(1011: ^18X1^1126 ^619*11 ^31^68化「化6 1^601510115 ^ 一 0 已 0 5 議 I 05015101^2 1^61 』0.0 0.0 ::;,1力 16.6

广

01^

1『^011 611161 005(5 !"311化「出出,1,『0厂"5 35 ^60151011 ^ ! ^ ! ^ , ^011 5110111^ 521601化6 -

1『3 ^1-0^1 0『1055 11!3^!\ 18 1101

01311011.

化爿,111611 3556551116111 15 ^0116 6乂 ^ ^ 6 1 * 3 ^ 8(^113!'6 ^!"^!" \? 1116

八5 11016^ 111 36011011 4.2.5, 1116 ^85&551116111: 1^1638111^6 13「0|3611^ 0311 1^6 561 10 1)00181011八、'61*386 8^1131-6 8「「0!", " ! ^ 棚 ! ! ! " " 。 ! ! , 01 " & . 丁I化060151011 0^11011 03115 「0「 1卜6 3556551116111 10 1?6 ^0116 6乂 1113X1111121113 口!"001 011 II化1)3515 0『1卜6 ^0^\1 611(61-6^ ^ 1116 1156匸丁 11;5 15 1111)51!'3^ 111 86011011 4.3.5, 丁卜6 1^180)118811100(1011 0「1161^011 15 111115^316111 36011011 4.3.7, 311^ 1116 ^ 6 1 * 3 ^ 8(1113^ 1^『0『0「!1611011 0「3550551110111 15 ^617101151^6^ !11 311 6X3111(316 1(1 8601100 4.3.8, -

八3 1116 0 醒 0 ^ 16^08 !1101*63863, 11化卩!"001 1113V 11101*6086 !0出&11 乂. ["!0\^6幫,1?6^011^ 501116 ^01111 1:1161*6 !5 00 3(^(111:101131 口「0?1|:國丁 1化^10^1 1113^ 丁卜6 130!III 31 ^卜」、1113「^1131, 0「 |1101~61116111:31 口"0^11 1)60011165 !1151^1111103111: 15 1116 0)3110111III 5126 0卩11化IV&&. ?

\\^1131: ^0110\^5 15 ?I 516^-1)^-51615 !1111511*31:1011 0「 3 !^^. 丁!16 1X66 15 8「0∼11 115!!!经1116 I!'31)11)1 ^ ^31:3 501 ^1111 71 52 !'000!"^5^ 丁 1化0『|3311:!1:1011 31^6 ^ ^ 0 | 3 6 0 1 311(1 1116 110^65 3「6 013551^6^ ^50 35 10 1113X101126 ^0^\\ 0!" 011化!" 4:「116「;011 01105611〉 1151115 ^ 1「3!1 "巧^313 561. 1 6

丁 116 乂311331;011 ^313 561 1156^ 「0!" 口11111!118 001151515 0『5364 「600「3^ 丁卜6 「1!165 0^ ^311:11:1011 311(1 1:116 110(16 013551^03110115 11131 1^6^610^ 011 1116 1「3;11!118 6^ 561 3「6 111611 3^116(1 10 311 5364 「^(;。!^^, III 6^601 1-6151-0^1101115讥化1^66 XV1111 1116 ^311^311011 ^313 561. 丁 1化110(^6 ^6^11 !110115 311(1 ( ^ ^ " ( ^ ( ^ 0「 1116 110(125 3「6 11化531116, 1)111 1110 1~600!'^5 1(1 63011 110^0 31^6 ^0111 1116 ^311^311011 6^ 561.

4.3.5

516|3~137~51:0|3 11111311-311011 0 1 01-0^71119 3 1 1 ^ ?!"11111119 3 丁『66

丁 116 131-00655 1)6^115 ^

1116 1713X1謹I 13*66 00118^010^ 「「0111 1116 !^!;!!!!^ 6^ 561


0^70^6^

4: 81411(^111^ 1)601810/1 7^66 ^0(^618

10

6(^1011168^01186 017^ /?/^

135

316|3 1八:焱1『66 18 3 6 ^ 6 | 0 口 6 3 1181119 ^ 6 ^31111119 3 3 ( 3 8 6 1

1^51118 1116 0161110(15 (! 13011536(1 III 6&1&\\ 3^0\6 出6

1111(1^1 11*66 15 66^0^66

9

1151113 ^ 6 11*31111113

3&1. 71115 11*66 13 3110^11 I I I ?!^!!^ 4.4,

1^0(16 1: 1^00! 110(16: " 。 , " ( ^ ^ 708, 1^00-1-651)011(161^ ^ 6444^ ?0181 ^ 7152. ?] 二 0.099, ?0 = 0.901 1*1-0(11 6601^100 I : 0.089, 1^0「11 !!"^ ^60151011 2 = 0 0^01510(11

111001110 ,' 525^

1^0116 2 ^ ^ ( ! ^ 二 251, 1^011-1-0813011(161^ : 2868, 丁0131二3119,?1 ^0.08, 0.92 , 「 『 。 「 1 1 00(16^601310(11 : -0, 12, ?1^0011111^61~ (^^"之二 0^ 0608810112

11081)011(161^ 457, 1^1011-!'681)011(161^ 3576, 丁0131 ^ 4033 「1 二 0, 113 , ?0 2 0.887, ?『00100(1扛(!沈^。" ^ 0.243, 1^0「11 3601510112 ^ 0, 0601810(11 3

3

主 八86 ,'

」,

540 35

540

35

1^0116 4 ^ 1x8(1: 81*0叩,) 116513011(10^5 ^ 100 " 。 ! ! - ! ^ " 。 ) ^ ― 338 70121 = 1438 ?\ ^0.07, 1*0二.93 060131011 ?『0「11 1 423 0 0 1)601810(12

10131 ^ 1681 -.91 ?\ :009,?0 ?『0「11 060151011 1 -0.01 0 0

1^0(16 6 ^ 1^681*3: 81*0叩3^ 1^51)011(16。 : 226 卜。。'『"(^"(^ 二 1944 70131 = 2170 ?I 二0」04,?0^ 896 060131011 卩『0「11 1 0.144 0 0

1^0(16 7 ( ! ^ " : ^ 叩 4 ^ 86515011(16!"5: 231 1^011-1^51)011(16(3 ~ 1632 70131 ^ 1863 ?1 二0124,?0^876 060181011 ?1-001 I 0.364 0 0

060|8:01|2

0601810111

1)601810111

1^0(16 5

&XI: 81*0叩2》

^ " " ( ^ :151 " 。 " ' ^ ^ " ( ! 。3=1530

^!2111*6 4.4 3110^5 1116 17-66 6&\^\0^6

&001 1116 / ^ ? / " / 0 1 0 1 0 . 丁 116 2^0\^ 11*66出&^"301 ^乂65

111^011113(1011 00: 1116 110(16 1(1611(1^0311011;出6

13611111103^011, 1^ 1116 110(16 18 3 16111111131 110(16; 1卜6

!!"!!!!^!^ 0『^?^^^^^! ? I I I 1116 110^6; 1116 !!"!!!!^!" 0『11011-?'65^0^6?'5; 1\\6 10X3.1 011011)0!" 0 『 ^ 6 0 0 ^ 3 10 -

63011 00(^6; 131*01)0111011 0严1*63口011(161*5 ^0816^10^

口"。!)"!)出矽0『「"^!!记);|5『。13011100 0『0011-

1-65^0110161*3 0)0516^10?' 口"。!^!)出杪0『!!^!!-^^^!!^^); ^0^\\ 0601310112; &1\6

1&1361 0『1116 360152011 [!1\61116

0001310111 ; ^0^11

131*0(115 3『6 111^X1111126(1『0「 1116 1\06&.

丁 116 11*66 0011515(8 0 ^ 1116 1*11163 0『36^10111005 0 ^ 1)16 !!^^^^. 31^111113『『001 1116『001110(^6 311(1 ^01113 0|0\^0 10 3 1^1^111101 "。^,

^011 0311 1*6^ 1116 01611111(1011 0『1116

110(16 0^ 1110

111636 36厂10!11005

&「6 3(3(6^1 I I I 1611118 0『1(16 111^111 !^!!径. 丁 1^6 111^11(3 3^160(6(1 5 ^ 出 6 11*66 31@0『1111111 I I I 1)113 6X31111316 5 6

31*6 38 ,化00016, 311^ 01*6(111 500『6, 丁 116『1^65 0『016^1111(10115 0『1)16 160/110^65 6

160/1:

1『111001116 〈 325,000

31*6

〈 35 乂631*5, 111611出6 !11(1^1011131『6001X1 ^ 0 6 5 10 ^ 0 " / ? 1

柯1、, 2: 1『1116化出乂&!!"、 11100016 〈 525,000 &1\6 ^ 6 2

35 ^631*5, 111611 1116 「6001X1 8

0 6 5

10 ^014^ 2


1\41^161': ?^001100\《。/"//。/^恤8148111655

尸 " ^ / / " / 化 ^ 4 0 ^ 6 1 1 ^ 讣 / / / ? & 4 5 ^6^156

^4^1100(10/75

3: I尸30 1 1 1 ^ 1 ^ 1 ( 1 1 0 0 0 1 1 1 6 ^ 525,000 3 0 3 0「6出1 5001*6 〈 540,出60 1116 1^600^ ^ 0 6 5 10 ^1-014^ 3

(^/;).

\^ 1\\& 11\61\\6\13\^5 111001116 15 ^ 525,000 311(1

160/4:

^?"。"/? 4 (^/

3001^6 ^ 540, 11160 1116 「6001^ @065 10

4\

3161^ 16:卩0316『10『|3^013313|11|168化『1116 163, 1100168什。171 1(16 1『3;。:。9加 \ ^ 卜 6 0 1116 131*561: 15 「651)0056 ( ^ ^ ! ! !5 3 1 ) 1 0 ^ 忪 『 8 0 , & 6

6

^0816^10^!)"0^1)11^3

3

31*6 1(16 !)'108出6

^1*0150111011 0『1*6515011(161^ &001 1116 13『0|3011100 0『000-1*65(300(161*8 111 63011 00(16, 1(1

^01)01)1111168 31*6 1156(1 35 ^1*6(110(10115 0 ^ 1116 !)!"。!^!)"^^^

1110^61 10 |3『05|3601 6111056 八11 005100161*5 (!^。。!^》10 3

31*6 3111*11)016(11卜6 8&1116 131*6(1101601 ^01)31)1111^ 0?『65130056.

8 ( 6 ^ 10: 0 3 3 8 1 ^ 0 3 ( 1 0 1 1 0,『6001X13 311(1 110(^68 ^

1113X1111123(1011 0, |3『0,115

013851^108(1011 0 ^ 1*6001x15 0!" 00(165 13 1116 0 3 0 1 6 8 ^ 6 1 1 10 1116 135化0『8331^111113 & I & ! ^ -

『6001X1八II 1*600^5 ^1111111 3 110(16 31*6 3531@116011116 83016 131^61 0『131*261 16乂61 10 &『6001"(! 0!" 110(^6 15 5^80(1 0 0 1)0111 1116 ^0516^10^

丁!16 ( ^ ^ ! ! ^01)01)111(168

61

1^61 10 3

10 3 5 3 ^ 。 & 01&33

( ㊀ 鄉 1 8 3^0^

31\6

1\\& ^0^\1 0131『1乂1八5 (163^I'1136^1 1泔6『,乂011 0311 1136 & 01*1161^00 0也6『11130 ^0^\1 1113X1111123(1011 10 0 1 3 3 3 1 ^ & 110(16 0『1116 1^600^3 ^1111111 & 0036, I ^111 11『51 ^61110113(1X1(6 1 1 0 ^ 013851110311011 15 3 0 0 6 1151113

01\1 \^\\&1\ 1116 131*361: 13『65|301156, 10 ?18111*6 4.4, 68011 1X0(^6 0011^1113 & 1^1)61: 1)601510/11 0『

0601810/12.八1^1)61

0『2)601510/11 016308 1^16 00(16 13 013831116^ 3 3 3 『 6 5 口 ( ^ ( ! ^ ^06&.八1&1)61 0 ^

1)601810912 0 1 6 8 0 5 1(16 0 0 ( 1 6 15 013551116(1 33 & ^\01\^&5^01\^

1\0^6^ 10 31X1^108 &1111636 ^ ! ^ ^ ! ! ^ , 1\\6

^0110^1112 口"。/"議(^乂 15 115601: 731)16 4.4

1 1

0 0

0

0

-51

八0001X1108 10 11115 ^0^11賺^IX,

1『1 001^601!)' 3551^11 1116 131^61 16乂61 1 ( ^ ^ ^ ! )

\0 3 11*116

&I ^ I X 11160 I 1118^6 & 510 口『。厂".1?1 100011*6011^ 3831@11 1116 1 3 ^ 6 1 1 ^ 6 1 1 10 & 0011-

^^!!^^!

1

『^口。!^", I ^111 1056 31 (& ^0^\1 0『一50, ?0「 6\3111口16, 10 110(16 4 1116 |30516『10『01*01331^111165 31*6 0.07『0『「65口0056 30(1 0.93 #0「 0 0 1^65^0118^ ^ ^ ( ^ ! ^ 2 ^ 0 ^ / 0 / 1 7 1116 6X^)6016(1 口『0111 ^ 0 . 0 7 * 8 1 0 十 0》3气一51 〉 ^ 一50,23, 1511^1)601510/12 00(16 4 35 3 ! ^ ! ! - 巧 ^ ! ^ ! ^ ^

^3.1 !5, 1^1 3601(16 10 3331^11

1^61 0, 0『013551^

1116 6X136016(1 ^0^1 ^ 0 . 0 7 * 0 + 0 . 9 3 * 0 ^ 50, 311106 50 ^ 一50.23, I

1 3 ^ 6 1)601510/12; 11x3.1 !5, I 0 1 3 5 5 1 ^ 00(16 4 &5 3 !!^!!-!^? !? 。!!^^!" 1\06&.八11『6001^5 I I I 3 110^6 31*0 33313116(1 1116 53016 131*361: 16乂61 51(106 1卜6乂 &11 1 1 ^ 6 1)16 53016 13031:6『10『131*0531)111(163 3 0 ( 1 1卜6 32111)6 ^0^11

0131『^, ?1^111-6 4.4 3110\^5 1116 ^0516^10^

^0^0^1111168『0『63011

^0110\^1115 03101113(10115 5 1 1 0 ^ 1 1 0 ^ 111636 ^03^10^^01)0^x111168^

\0^^^\

110(^6 &5 ? 1 311^ ?0^ 丁116 \ ^ 出 1 ) 1 6 /?/"^///議IX, 31*6

1136(1 10 & 5 5 1 ^ 1 63011 0 0 ^ 6 10 & 131*^61: 16乂6乜 ^ 0 ^ 6 1 ( 『 。 ( ^ !!!^) 1 1 0 0 1 6 1 - ^ / 5 / 0 ^ 7 1116 6 X ^ ( ^ 6 0 1 ^ 0 0 1 ;3 ^ 5 1 0 * 0 . 0 9 9 十 ( " ^ 0 * 0 . 9 0 1 ^ 50,089, 0 0 0 1 6 1 - ^ / 5 / 0 / 7 2 1116 6X13601601 「0『11: 13 ^ 0, 311106 1(16 ^1*0^1 111136『06(^8^11 !5 0601510?12 0601810111 9

!5

V 111311 1116 ^10^\\

丁卜31 !5, 1(113 00(^6 15 35513116(1 1 0 出 6 13^61:

1:1115 0 0 ( ^ 6 &1*6 013551^16(1 3 8 巧 ^ ! ^ ! ^ .

1.八11 016011)61-5 0 ^


017006^

漏 6

4: 8""出1^

060151017 7^66 1^0^615 10尸/^/亡/

^68^)0)186 011^ /?/^

137

2 1116 6X^601601 口「0^ 15 51 0 * 0 . 0 8 十 ( 一 》 1 ^ 0 . 9 2 : 一 50, 12,

1 1 0 ^ 1)6018X0^1

6X156016^ 131*0111 15 0 * 0 . 0 8 十 0 * 0 . 9 2 ^ 50^ 8 1 0 0 0 1116 ^0^\\ \\^^ "!^"

06^0111

漏 6

3

, 11115 110^6 15 3551^163 10 1116

1)601310112

0601510112 1116

1116 ^1*0111

!5 ^6316^出30

\^\^\ 0 ^ 八 I I 016011)61-3 0 ^ 11118 00(^6 31*0

1100161- 0601510171 1116 6X^0016(1 口『0^ 13 : 810*0.113十(一5 0 * 0 . 8 8 7 : 50.243^ 1100161- 0601810112 出6 6X130016^ ^0^\1 !5 0*0.1 1 3 十 0 * 0 . 8 8 7 ^ 50, 311106 1116 131-0111: 110016「 1^1310^1 13 11130 1116 ^0^\1

11036「 06(^810112, 11113 110(16 !5 &331@11001出6 13^61 1^01 1 ,八II 0161111?6^5 0『出;5 00^6 31*6

013851116(1 35 !^^!^!^. ^0116

4 0601810111

1*16 6X^)6016(1 口『0^ 15 : 5 1 0 * 0 . 0 7 十 0 * 0 . 9 3 ^ 一50,23,

0601810112 1116

6X^56016(1 |31^0111 15 0 * 0 . 0 7 十 0 * 0 . 9 3 二 50, 3 1 0 0 6 1116 131-0^1 " ! ^ ^ 1 ) ^ 7 5 / 0 ^ 2 !5 "!^"

111311 1116 ^0(11

06(^10?! 1 , 11115 0 0 ^ 0 13 3 5 5 & 0 6 ( 1 1116 131^61 16^61 0,八II 01601(361-3 0『1)115 110^6 31*0 013531116(1

35 !^!!-!^^!!^!^.

0601510/11出6

6X^)6016(1 口「001 !5 = 51 0 * 0 . 0 9 十 ( 一 》 1 ^0.91 ^ ―50.01 , \31\661 0601310112 1116

6X136016(1 口『0(11 13 0 * 0 . 0 9 十 0 * 0 . 9 1 ^ 50, 81(106 1卜6 口「0闩1 110016『06(^10112 \5 ^ 1100161- 0601810111^出15

&

0 1110 ^1*0^1

00(16 15 3551^16(1 1116 1 3 ^ 6 1 1 ^ 6 1 0,八11 016011)61*5 0『11115 110(16 31*6 013551116(1

35 !^!!-!^!? 。!^!^. 漏 6

6

!^&!" 0601510121 1116 6X^)601001

|5『0111 13 ^ 510^0^ 1 0 4 十 ( 一 》 1 ^0^96

144, 1100161- 0601510112

^

1116 6X136016(1 ^0^11 15 0*0^ 1 0 4 十 0 * 0 . 8 9 6 ^ 50, 81006 1116 ^1-0^11 11110161- 1)2018101X1 15 卩『0厂11 11036「 。601^0172, 11115 110(16 13 3531@11601 1116

01&551116^

111311 1116

\^\^\ 1 ^ 八 I I 016011)61-5 0『1(115 00016 31*6

28『^(^(!^.

7 0601510111

1116 6X^56010(1 口『0厂11 13 : 5 1 0 * 0 . 1 2 4 十 0 * 0 . 8 7 6 : 5 0 3 6 《 & V

1:116 6 X ^ 0 0 1 6 ^ ^51*0111: 13 0 * 0 . 1 2 4 十 0 * 0 . 8 7 6 : 50, 311106 1116 131*0^11 ^0^\1

[!0601810^2^

1)601510^1

0601510712

15 8「6&16「 11130 1116

0『11115 110(16 31*6

1(113 !10^6 !5 35313116(1 1116 131^61 1 ^ 6 1 1 ^八11

013531116^ 35『^^。(^(^!^.

3 1 6 ^ 2: 0310111311011 0^ ^311^311011 |3『0门15 311^ 3616011011 0^ 1116 1^9^*3:260111*66 ( 口 ∼ ! ! ̶ ) ?1『51, 1116 110^6 ^11X1110115 31*6 1156^ 10 |5311!11011 1116 ^311^3(10(1 00(16 15 311-63(1^ 35512116(1 & 13「8

030 0310111316 1116

^0^\\

61

0『63011 00(16 0『1116 11*66 1131113 ^

1)1*0111:3 31*6 0310111310(1 6*001 1(16

1010山伃61*601 !!。^^? . 311106 63011

1)356(1 0 0 1116 ^0816^10^ ^01)0^1111168 1 6

3001 1116 ^0^\\ 01311*1 \

^^11^(1011 6^

5&1.

31006 1)16

1(^3(1011 6&X& 56仁1卜6乂 31*6 0&116(1 ^011(^011011 / ? " !

5110^5 1116 3^110311011 0『1)16 11*66 1 0 出 6 ^011^0(1011

561.

7

! ^ ^

4.5

9

I


138

?1-6(^1011^6 140^&11^

V61": ?^001\00180114(10^3 ^

848 ^16^186

8143117655卸口 11(^10173

1^0(16 1: 8001110(16: ( ^ ( ^ ( ! 饥 2 560, 1^00-116513011(1^5 4804, !"0181 ^ 5164. ?\ 二 ^). 104, 0.896 06018100 I《18巧"1(^1 。 !) 3

111001116 ,' 525^

1^0(16 3 ^ 网 廿 加 358, ?^!!-!^网^!"5 ^ 2690 70131 二3048, ?] ^0.117,1*0 = 0.883 0^151011 1 ( ( ^ ! ^ 1(^1 !)

1^0110 2 ^ ^ ( ! 奶 202, ^011*1-681)011(16!'5 二 21 14 丁0131 = 2316, ?1 ^ 0.087 ,1*0 = 0.913 06015:0112 ( ! ^ " 1改6, 0》

3

3

2

3

丄 ^^11 500化 上

,'

35

540

540

35

1^0(16 4(1^!'1:8 加 叩 !)

^ " " ( ^ 二 72 :1004 70131 = 1076 ?\ : 0.067 ,?0: ,933 ^"'『。(^(^

0608(0112 ((&1^61 !6^61 0》

1^0^16 5 (!^&12 : 81*0叩2》 ^""(!饥:130 !^-战!^)!^" 二 1110 70131 = 1240 ?1 ^0.105, ?0二.895

010||6 6 ( ! ^ ^ : 81*0叩3^ 1^51)011(16。 : 177 1^0051)011(16!"5 二 1448 70131 = 1625 ?I ^0.109, ?0二.891

06^1510112 0。巧"16^61 0》

06015(0111 (化巧"16^61 !)

八&6『3^1^1113 1116 1)06& ^6110111008 1 0 出 6 0001^&^108 1116 11*66 & 0 0 1 1116 ( ? ! ^ ^ 4.4〉, 001100

00

1(^311011 6&1& 561, I

^0(16 7 ( ! ^ ^ 4: 81*0叩4》 ^ ( ^ ) ! ! ^ ! ^ ~ 181 "。!!-!^""!^ : 1242 70131 ^ 1423 ?1 : 0.127,^0\873 060:5!01|1

1^16 11*66 5110^0 111 ?1^1『6 4.5,

& ( ? ! ^ ^ 4.5》\^\\\\ 1^6 1『66 & 0 0 1 1116 ^1111113 ^ 3 ( 3

1116 0(60151011 135613 [!)6^510911 0『1^!5!0"!) 0^ 630)1 110(^6 &『6 6 X 3 0 1 1 ^ 1116

531116 111 1)0111 ^ 1 6 (!!&@^1115^ 丁 ^ 5 000111-5 1)6031136 111036 ^601510113 31*6 6 3 5 6 ^ 0 0 1116 110^6 ^6^111(10115 80(1 1卜61)08(6^10^『(?!)^出"63 361161*3(6(1 ^111*1113

11*31111113 0『1卜6 11*66 (^^!!!^ 1)16 11*31111113

561〉, 丁 116^ 1*6001116 13311: 0 ^ 1116 1X10(161 311(1 ^ 0 0 0 1 01181^6 \^|1611 1116^ &!"6 3 口 1 0 & 0 6 ^ 6&10.

1131113 111656 110(56 ^6^1111(10113 3.1X6 11115 0135311103(1011 0 『 ^ 6 110(165, I ^111 0 0 ^ ^61110113(1-3(6 |10\^ &11 0(3(111131 11*66 13 5 6 | 6 0 1 ^ 7116 11-66 111

4.5 1 1 3 3 『 0 0 『 1 1 0 ( ^ 8 、 311(1 I V&^ 10 1^\\8 11*66 35 1116 1110X111X01 ^66^ 1^0^6^6^

^1(11111 11113 11*66 11161*6 31*6 56乂6『31 5111)41*665 0『311?6『6111 51265. 71161*6 31*6 1^0 ^665 0『5126 3, 丁II&I 13

9

1^1 口『11116 110(163 6 &03 7, I XVIII @61: 3 3116~^66 "!出110(163 3, 4, 311(1 5 3 5 1 乜 1 6 ^ 6 8 ^ I ^111 0&11 50(3-11*66—3.1, 111(1103(1113 I 1)1*11116 110(163 4 311(1 5, I ^ 16^68^

I

& ^135 1卜『66 1 6 ^ 6 5 30(1 11X3.1 I I !5 1卜6 ^11*51 0 0 6 0尸1116 1111*66-163^ 11-665. 1『 31101116『?!!!?-打枕0^ 5126 3, 11113 111116∼&卜110(165 2, 6, 311(1 7 33 1(5

0&11 11113 511^-仕66^3.2,八11086^6!", 11161*6 31*6 ^ 0 8116-11*665 0『5126 3, ?『001111636 1\^0,

I 011151 3616011(16 0 0 6 1卜&1 ^161(15 1^16 1 1 1 ^ 1 6 『 3 1 0 0 6 1116 ^0^\1 18 03101113(6(1 1151113 1卜6 1(13(1011 6&1& 561, 11 15 03\\&6 ^011^011011 ^0^11 111 6 0 1 6 ̶ 5 6 1^111161\ 10 1116出501133|011 1?610\^ I 1136 1116 161*1115 ^011(^0(1017 ^0^\1 &0(1 ^0^11 10(61*01131186313 !)'.


01X00&V 4: 81411(^1^ 1)60151017 7^66 \1。^3 10 ?^^^1011^65^01156 0^

139

丁116 116X1 3 ( 6 ^ 13 10 6X31111116 11)6 |5『0111 0『63011 /^?/"&001 31|13^66^3^ 1 ^ 1^0^6 3 1135 358『65口00(161*5

2690

0111 1116 ^311(1311011 ^ 3 1 3 561〉, &1\6

000-1-6515011^01^

110(16 ^ 3 5 &11*63(1^ 01&851^16(1 &3 3 "鄉。"^

5890

3 0 11 1135 8

561扒81006 ^ 1 5

1^(1011 151*0(11 0『

3 5 8 * 5 1 0 十 ^ ^ ^ ^ ( ^ - ! ) ) . ^ 0 ^ 6 4, 0 0 1116 0出6『1\&1\6 13 & 。 。 。 - 『 " ^ " ( ^ ! " 00(16, 9

72 # 5 0 十 1 0 0 4 * 5 0 》 1 ^ 0 ^ 6 5 13 &150 3 !!^!!-!"^^)!^^ 00(16,

1(13(1011 ^0^\1 0『50

&1\6 50, 11 &150 ^161(13 3 口『。份。『50^ //(?"(:6, ^6 1010,1卩"//^/。"。" 5145 ^66^3.1

1(1^1011 6^

2 ^ 6 1 1 2^00\6 10 0^101113(6 1卜6 口「0111. 丁 116『6『0『6, 1116

0011111111 0 ^ 016 ^0^\1 出15 163^13

0111 1116

?10^6 ( ( ^ 『 ! ! ^ 11*31 I I ! ^ ^ ) , I 1111151 1 1 3 6 出 6 ( ! ^ !

^0^11 0^^11出6

160^65

0^

18 5890, 1^0^ I 03101113^6 1^16 10\^\ ^0^11 0^50卜汁66一5,2, ^ 1 0 1 1 1133 110(163 2, 6, 311(1 7

&3 1(8 160^63^ 311106 110^6 2 15 。1353^601 3 8 3 11011-1*65口。!^^! 110(16, 1(3 |31^0^11 13 60)1131 10 80 1151113 1116 1

360011(1 0011111111 0 ^ 1^16 口『0111 0 1 3 ^ 1 \ 1^10^6 6 15 &『^^「(^(^! 110(16; \11135177

1*65口011361*5 30(1 1448

1

1

000-1*631)011(161-5 111 ^6 \0\\^01101\ ^010 561. 5 。 1(8 10(31 口"。/" ^ 1 7 7 * 5 1 0 十 ^ ^ ^ ^ - ^ ^ ^

5322,

1^0(16 7 13 3150 3 1*65口。口^ 1\06&. II1135 1 8 1 『 邵 。 ! ^ 。 &0(1 1242 000-!"65|3000|61^ 丁01^1 ^0^11 0『 ^ ^ " 。 " : 181气510〉十1242气一51〉25568^ 7 0 / 0 ; 0 ^ 5 1 4 ^ 6 6 ^ 3 . 2 30, 1 ) 6 ^ 6 6 0 1116 1 ^ 0 ?!!!)"^^^ 0? 3126 3

15 5890

^322十5565^

11^6 3 116 III 161*015 0『1116^ 口!"。^!^.

1^0\^ I 01051 6X31111(16 &11 3 1 1 ^ 1 ^ 6 5 0 『 1 . 10 1116 6X3111 口 16 11161*6 13 0 0 ~ 0 0 6 511011 5116,1『66, 311(1 I ^111 0311 11 811^-^66^.!.

I I 1133 0 0 ! ) ' 1 ^ 0 163?00365, 030161 乂 00(16 2 &0(1 0 0 3 6 3, 1^0(10 2 15

013551110(1 &3 3 !!^门-巧^口^!!^^!^ 110^6; 1161106 11 ^!11 ^161(1 3 口 1*0(11 0『26『0, !^0(16 3 13 01383|116^ 2 8 & 358 1*65口01^61*8 &0(1 2 6 9 0 !!。!!-!^^)!^!^ & 0 0 1 出 6 ^ 1 1 ( ^ 0 1 1

^65^01^6^11\0^\11135

力61(15 &/?―/ : 358吖灣十2690*(-3】):5观7116101&1 !^0∼,1^1: 113 03101118(6出6 ^0^11 0 ^ 1^6 ^001 "(^.

& 561. 71113

口 「 。 份 。 " 油 - ^ 」 . 1 :5 5890^ 丁 116『001 110016 13 018851(16^ 3 5 3『^^。(^(^『,&11(1 !1

1133 560 1*63口0110161*8 311(1 4804 000-1*6513011(16!'8 111 1卜6 乂311!011 6^

5

6

3

^0^11 0^5796,

I 3110111(1 3130 0310111^(6出6 \011^011011 ^0^11 0『1卜6 1113X111131 1『66, ^111011 13 1卜6 1^66 "!出『011『163^ 。。(^?. I 0&11 11113 31115-11-66^4.1 ^ 丁 1^5 1?66

3 5 110(165 4, 5, 6, 311(1 7 &3 118 1 6 ^ 6 5 ^ 1^0(168 4 311(1 5 &!"6

013351^1601 3 5 11011画『65|301^6『110^65 311(1 1161106 ^161(1 261*0

1^1*0111 1116 03(01113(1011 3^0\6

9

9

110^6

6 15 3『65|30036『0036 30(1 乂 161(^5 & |3「0厂"0^5322, \^|1116 000)6 7, &150 3 1*651)011016!" 00^16, 乂 1 6 ^ 5 & ^0^\1

0『3568, ^ " ! ! 径 10(^1 口『0厂"0^ 5890『0『出6 1118X111131加

73616 4.5

3

3110\^3

1116

311111111317 0 『 1 ^ 1 6 卩 " / / ^ ? / / 。 " ^0^1(5

0『&II ^!!^-!^^ 311(1

1116

11*66:

791)16 4.5 811111111317 0\ ^31111311011

! ^^ 3

巾136「

1 2 3 3 4

尺001 00016 5013^66^2.1

\^3||0|31100

0, 163^63

丁「66 0 0 0 1 6 3 2 30(1 3》

3 1 1 1 ) ^ 6 6 ^ 3 . 1 ^ 0 0 ^ 6 3 3,4 30(1 5》 811^11-66^3.2 ( 门 。 ^ 2,6 311(1 7》 3 1 ^ 」 「 6 6 一 4 . 1 ( 门 。 ^ 4,5,6, 311(1 7》

八 5 0311 1)6 5 6 6 0 &0111 1116 31?0乂6 13^)11\\& 0『1116 1 ^ 6 3 ^ 画 6 《 尸 0 ^ 16(^68

3796 3890 3890 3890 3890

0)3(111131 1 ^ 6 !5 50134『66一2, 1 ^1111 1 ^ 0 163乂65, 66031136 &II

6^11311^出呂卜131*01115 31*6 1^665 0『131^6『5!2巳八5 ^011 @0 ^ 0 ^ 0 1116 00111171115 2111(1 「"//^?//。"卩"。/^, 110^106

&5 1)16 0腿1?6『0『16^68 1110^3803

9

11161*6 13 0 0

11101*6380 111 |5『0厂 11. 5(5, ^I10(116I'

10 111111^ 0『11113 00111(3&1*15011 " ! 出 ^ 0 6 3

1)6X011(1

^0^11 ^ 0, 71131: !5, 1^16 11101*6111611(31 ^0^\1 |36『3(1(11(101131 163^ 15

2.1, 1\\& 1^1^1^9.1

261-0. 1-16006 511^11 66一2, 1 13 1卜6 0^10111111 0 『 ^ ^ 5 1 2 6 ^ 1

3 3 出 6 1113X111131 1『66^

5

1『66, 1^016

!3 10 58^

31115-11-66^4.1 15 1116 3&1116


?1-6^10^6

4.3.6

^0^61117^

^[!1761': ?^001\00\ 8011(110178 /。圹8148111688 ^^1100110118

" 〃 / ? 8/18 ^6^136

^61*396

^1*0181習5, 丁 0 1 3 1 ? ! ^ ! ! ! 10『00111|131 8|19 丁 『 6 6 5 0 1 |

08^61-0111 8 1 X 6 8 0X3171)310 I 0^101113(6^出6 ^0^\1 0【& 11*66 〈0『?^^-^^) 3 5 1116 31101 0『1116 ^1*0^118 0『1(3

10 1116 2^0^

160/170368, 丁 116 ^0^11

60)11&15 10(01 ^0^11

『0『&11 1116

5364 1*6001^5, 0^^111011 560 &1*6 ^ ? ^ ! ! ^ ^ ! ? . 丁116『6『0『6 6^6^6 ^ ! ^ ! ^ \^ 5 3 6 4 811106 1116 36000110&10『(!^!!!!^!" 0^!^。。!^^) 13 1116 3&1116

1(13(1011

& 1133

-

001X^3!"63, 1*6831X11655 0^出6汁011011)6!" 0 ^ 16(^63^ I 31X1^6 3.11116 531116 0^(111131 11*66

1131113 6!1116『0^6^0^6 0『10(01 ^0^118『0『0001)331^500, 110(163 15 6^11^1 10 1)16 10(31 I I 讓 0 ^

I & 6 ∼ 1 5 6 , 1116 5111X1 0『1*651300(101-5 111 311 163^

1*651)011(161*5 111伤6

6X3111 口 16. 961106 03101113(1113 0^6^0^6卩"。/"

1(13(1011 6 ^

&5 10(01 ^0^111

561,

15 560 111 011*5

560『0『63011 ""!?-^^!"66 ^111 131-0(11106

1116 53016 0|51111131仃66,

4.3.7厶。0111*307/1^50隱3550318011

0^1161*1011 811 8616011119

1116

1^19^

812611 丁『66《0画3551,8。311011 。詈11600『115 31111 ^ 1 0 ^ 6 3 \ ^ 1113X1111121119 丁 11115 &『,I 1 1 ^ 6 016501*11)6(1 110\^ 10 01338117

6

^

0 0 ^ 6 &1\6 \\0^ 10 ^^176 3 ^ 6 6 1151113 ^0^\1

1113X111112&11011 3 8 1卜6 0『116『!00,八00出6『0^611-1156(1 0『116『!00『0!" 1)0^1 0 0 ( ^ 6 0133811108(1011 811(1 11*66 ^1*11111113 15 ^ / / " " / / ( ? " 00011^00^^ ^\\\0\\ 13 ( ^ 。 『 ^ ( ! 1 ) 6 1 0 1 10 01"(!6!" 1 。 1156 11115 0『116「100『0『501311-66 361601:1011 0『131*11111113 111 5 0 1 6 ^ 1 3 6 ^4;116『,乂011 1111131: 3 6 1 也 6 ^886551116111 ^ 6 3 8 1 1 1 * 6 |31^0|3611^ 10 1\11801388100&1100. 8 1 0 0 6 111150135511103(1011 1*316 ^ 1 一 ^311(1&1!011 30011『30乂, I 1136 1116 1611113 ^011(^011011 000141-00^ 3.1X6 1^71501(3551/100^1011 11^61*011&08 化1》 6

\ \ ^ 6 0 1116 300111*30^ 0『!16『100 15 3616010(1『0『35565501601 0『1卜6 11*663 10 018831^16(1 ( 鄉 ― 6 。 8

&&000^6\1\^

1*5 30116 ^111*1113 ^ 6 11X11111113 0卩出6 11*66 1131113

4.3.7.1

10 1113X111111111 ^0&16^10^ ^01)01)1111^^ 71115 3 5 5 ^ 細 6 1 1 1 ^111103 (!&^

0338|||0311011 0! 1116 1^10^68 \ ^ 1\13\:11111111 ^08161 101 1

丁 1115 0 3 0 1)6

1110118111:

0『35

1^16 110(163 31*6

|

^0011^30^

1151(13 3 ^0^\1 0 1 & 0 『 1 1 1 6 ^0110^102 '!36:

73(316 4.6 厶0011巧0乂 000131011 1

0

1

1

0

0

0

1

10 11115 11131『1乂, 1『& I^6500I1^I6「 !5 001^6011^ 01&531^16(1 3 3 3 11*116 ^^^^!!^^!', 11160 0 0 6 111111: 0『300111*30^ 15 3 ; 0 6 丄 1 『 & 000-『65150036『15 00176011^ 013551116(1 &5 。。!!-!^!?。!^!", 11160 0 0 6 11011 0 ^ 30011^30^ 13径&!!^丄01)161^156 1116「6 15 0 0 ^ & ^ . 八5 5 ^ 0 ^

5

1:116 1100165 3「6 013331116(1 & 5 化 ^ ) 。 ! ! ^ 0「 000-『631300016「 110(103 6830(1 0 0 1116 ^051:6^101

-

131-01)31)1111163 0^101113(6(1 1^0111 1116 / / ^ / " / ^ ^0(0 861 2111(1 1卜6 1116 II &101112 -

90.1

361 1116 ^ 0 ) 3 0 1 1 1 0 0 0『『6印000|6『5 13 9.9^0,

^0^\1 ^^\\^

1(1 1116 1*001 0 0 ^ 6 0『

1\\6 1^01)011100 0『000-『65|3000161^ 13

?16006, 1『1116 1-001 00(56 13 013531116(1 3 5 & !"65口011016『00(^6,

30001^30^ ^011101 5 6 9 . 9 ^ ; 1^ \1

15 013331^16(1 3 5 !!。!!-!^^!^。!!^^!" 00(16, 1116 30001^30^ ^0111(1 1)6 9 0 . 1 ^ , 9 6 0 0 6 , 1^16 0^1111^1 0160131011 13 10 0 1 3 8 3 1 ^ 1)16 1*001 00(16 3 5 3 !!。!!-!^!^!!^ 1\06&.


^002^

4: 814脑"8

060181011 7^66 1^0^18

81011131*1 乂, 110(165 2, 3, 4, 5, 6, &1)6 7 ^111 311

I。

101 ^5/70/256 0^

1^18^ 1 4 1

。1855^3 35 。。!!-!^!^)!!^ 00(165, 511106 1116乂 311

0101-6 !!^!!-^? ^^!!^^!'? 111311 「65^3000161^

4 . 3 . 7 . 2 03101113(1011 011116 1^:30|388:1:031:071 ^316/^00111*307

311^ 36160(1011 0于

1116 1^19111-3126^ 丁『66 31006 1116 011501&551110311011 1*316 15 1 ~

1(1311011 300111^30^ 111^X1111121113 1116

1^^1011 3001)1*30^ 15

5

1116 53016 &5 1111111111121113 1116

「&16.

丁31)16 4.7 5 1 1 0 ^ 5 1116 300111*30^ 0 ^ 1116 ^1061^6111 5116-1!'663 \^|1611 3|3|3||6^ 10 1116 \01\^01\011 ^0(0 115108 「0165 30(1 138111(10111113 ^11111(10115 ^ ! - ^ ^ ^ 『 「 0 0 1 1116 //^/"//^ ^0(0 861, 811 1^16 0 0 ( ^ 5 31*0 01355|1 16^ 3 5 000-「65|300016『00^63^ 丁 116 ^ 6 6 1135 1111*66 311|3~^665 (!!。^ 0011011113 1116 1*001 00(16 3 5 3 :

11-60)1 51164『66一3.1 ( ^ ! ^ 00(^63 3, 4, 30(1 5〉, 511(3-11*66—3.2 ( ^ ! ^ 110(163 2, 6, 3X1^ !), &1\6 5\1^~ 11*66—2, 1 (^!!!! 110(165 2 311^ 3》丁卜6 ^011^0(1011 00014^0^

0『5116-1「66一3, 1 ^ 〈 2 6 9 0 十 1 0 0 4 十

1110)75364 ^ 8 9 . 6 ^ , &1\61\\& 0115013531^1031100 1*316 13 10.40/0. 丁 116 ^011^011011000141-00^ 0『511511^66—3.2 ^ ^2114 ^ 1448十1242乂5364 ^ 89.60/0

1116 001X65(3011(110& ( ^ ! ^ ^ ! ^ ! ! ! ^ ^ ! ^ ! ! 『 & 忱 1 5

0『50卜11^6^2, 1 ^【21 14十2690)75364 ^ 89,6 人丁卜6 \01\^01\01\

10.^/0, 7116 ^011^0110^1 00014^

0

000141-00^ 0 ^ 1^16『001 110(16 13 ^^/^^(^)

: 8 9 . 6 ^ , &0(1 1116 11115013531^103(1011 I &16 15 1 0 4 ^ ^ 丁 116 -

^011(10(1011 000141-00^ 0卩1116 018X111181 11*66 ( ^ ! ^ 00(165 4, 5, 6, 31101 7〉 ^ 〈1004十1 1 1 0 十 1 4 4 8 十 1242)75364 ^ 89.67。

3 11113013551110^10111*316 0『10.4^0, 丁 116『6 15 0 0 3311110 30001-30^ I?乂 ^

0

'

0

^

1)6^011(1 1116 1-001 1106&. !"16006, 1116 0^(111131 0『1*18111-5126(1 11*66 15 & 11*66 ^1111 1116 1*001 0 0 ( 1 6 3 5 1(3 0 。 1 乂

731)16 4.7 311111111317 。, ^311(1311011八0。11『30乂 闪訓136「 0《163^5

丁「66

? 「。汛

1

89.60/0

501)^66^2.1 (^^^^^ 2 31101 3^

2

89.60/0

5 1 1 6 ^ 6 6 ^ 3 . 1 ^ 1100165 3,4 3001 5》

3

89.60/0

311^1-66^3.2 ( 门 。 ^

2,6 3^6 !)

3 0 1 ) ^ 6 6 ^ 4 . 1 ( ^ 的 4 , 5 , 6 , 311(1 7》

4 . 3 . 8

^836331116111 。 1 3

3 4

89.670

01 ^ " ! ^ ! ^ 1

1181119 ^ 6 1 * 3 9 6

^ " ^ ! ^

1116 1 3 ^ 6 1 13 000110001)5,出6 3536551116111: 0『& 11*66 15 ^ 0 0 6 1151113 511013 0『3^1131^6^ 6 汀 0 「 ^ 八 5 111 1)16 0 3 3 6 0 ^ & 1)10317 13『〇61, I 1156 1116 \011^01\01\ ^0(0 82110 1301^01111 1116 &556551116111. 丁 116「6卩0「6, I 60^1(1 \^ 3(3)31^1113

110(16

111(10115 0「 「11165 0『^21111110111118 ^ 6 ^ 0 1 0 ^ 6 0 1 1181113 1116 (!1111113

5 6 1 1 0 1116 ^3113311011 63.1& 561. 丁 116 \^01111 0 『 3 110^6 15 0 1 6 3 5 0 1 ^ 3 (!!^^!^^乂) !)乂 1116 \ \ ^ ^ 3 ^ 0 3 1 ^ 6 1 ", 6汀0『『0「 11131 0 0 ^ 6 ^ ?0「出6 ('^ 1\0^&, 1116八乂6「386 3(^1131*0 6 汀 0 『 ^ 一 V ( 乂 , 一 又 〉

1116 301:1131 ^3(116 0『1116 131*^61 ^31*131)16化!" 1)16卢'『6001X1 111 1116 1^ 00(16,

2

,∼!161*6 乂, 13

歹 , 1 3 出 6 6X^601601

^ 1 1 1 6 0『1(16 131-261『0『1(16 /'力1\06&. 丁 116 6X136016^ ^ 1 1 1 6 义 1 5 021101113(6(1 &。01 1116 //^/"/"^ 63.13. 561, 1)1111116 01)501^3110115

31*6 & 0 0 1 1116 \011^01101\ 6&I&

& 0 0 1 1116 11-0111111^ 6&1& 561.

丁 116 00(16 ^611111(10(15 &『6 0005(1*1101601


? ) ' 她 ^ ? 〃 " 《 1 1 7 7 / 1 548 ^11(61-^)115^入41贈:?!'^!!"!!

80(1(1(0^力!' 8115!醫5 /!/?/?/!0^110/15

丁 116 0^61-311八\^「386 5^|113「6 0「0「卩0「 3 1^66 0「 5116-1!"66 15 1116∼^^]^^! 51^ 113^6(1 61-1-01-5 0!" 1)16 111^1^1^1131 163乂65, 311^∼扒&卩0311 1?6 35 ^0110^/5:

0『1卜6 3乂6「3〇6

03551?103!!011 0『110^65 15 1101 11000553^『0「 0 3 1 ( ^ 1 & 1 1 1 6 3556851116111 ^ 1 1 6 1151118八\^「3^6 5(^1131^已「「0「 (!^。^"^^ 1116 1^01 III 1)115 0356 15 ( ^ " " 川 。 ! ^ ) .

4.3.9

3 6 1 6 0 1 1 0 1 1 0 1 1^16 1^111-512611 1*1*66

015(313^ 4.7 21^65 31101116!" 111115(1*311011 0卩1卜6 5616011011 0^ 1116 ^1^11-5126^ 1^6&『「0111 3 50(^1)61106 0『 0(3110121 (!"663, 1(1 1116 ^ 口 " 5110—11 III 0!3口 1 巧 4 . 7 , 出 6 110 「;2011^1 3X15 3|10\^5出6 。 醫 ^ 0『16^63 111 03011 01116 1:!"665 !11 1116 0|:1;111131 86(^1161106 0 ? 仃 ^ . 丁 1化^61^1031 3X15 5110*5 1116 3^61*3^ 0「0厂 11 0『 03011 11-66^ ?1~0111 1116 3;3卩!&^ !1 !5 0\V 1)131出6 0|31111131 11*66 1135 86^00 1@^65 71131 !3, 3 11*66化1出 56^0 110(^65卜35 1116 531116 1(^91 !011 ^0(11 35 311 丫 1^6 11101^6 110^65^ 1461106 1116 ! 1 - 6 6 5 6 ^ 6 1 1 1^3^65 15 50160(6(1 ^出6 31^01!III111 (^^^ 1116 ^6111031 II116 。0「1^59011出118 10 56^611 16^65 011 1110 1101*12011131 1

-

0(51313^ 4.7 3110^5 1116 111111311^ 001151 「11(^6(1 013X111131 ^66 0『5126 14, 136『0「6 [^「"门""^ 11 10 3 11*66 0『 5126 7 115102 "化【^川!^!化^650!'11)6^ 631.1^ 018 口 13乂 4.7 ::」崩

丁 10|【1: ^171^1^ ^11^110「


謂4焱060画3,011

7 『 6 6 11101101 1 0 ? ! ^ ! ^ 量 1 ( 0 3 ^ 0 1 1 3 6 1 0 0 画 『 6 。 1

|

1\1131 |101||19 III 11113 36011011 I (!^^^^^ 3 1^5|?01156 ^0^6\ 1131113出6 1)6018100 7|^66 110016. 1116 1110(161 13 1)35601 0 0 ?

311111113(601 6&X&『6口『656!1^113 1116 6\1601 111311 03111^1311 0 ^ 3 11^^0111611031 1115111^1100 00111口311》丁卜6 |311175056 0『11113 3601100 15 10 5 1 1 0 ^ 110\^ 10 ^6^610(3 & 1*63^01136 1110361 1151113 1卜6 1)6018100 1 1 * 6 6 00(16, 3 0 3 10 11131(6 乂00 & ∼ ! 出 1 1 1 6 乂3『!005 00111(3011611(5 0『1^16 1^68111(8 "^(!。∼.

8 1 0 0 6 1)115 010(161 ^6(11015 311 1(1(11^1(111&^3 !'65^0056 10 1(111113 010(161, 1116 I & ^ 3 1 * 1 3 5 1 6 0116 ^31*131)16出&11

冚&出!^, 11 13 0^\\^6 &

^65^701156動

10 ^!"^!^) 13 1)111317: I I ^ 6 3 0 0 乂31065

0『^651)01136 311(1 710 ^68^)0118

丁 116 口『00655 ^ 0 \ ^ 011^1^111『0「 111!5 1110(161 15 5 1 1 0 ^ 0 1(1 0|5|31^ 4.8^

018(1137 4.8

^53?

0813 ?3|1111011

0801810(1 丁『88

10 1110 1)1-00653 ^ 0 \ ^ 41&径『&卬1^6 ^11^1 0 0 ( 1 6 !5出6 1111)1111)9(3 ( ^ ( ! ^ . 丁116 ^^0^611168 0 『 ^ ! 5 110(^0 &1*6 5 | 1 0 则 I I I 013^)13^ 4.9^


144

&7/1:'/'/?/'/.化^!化V: ?1~0011001 8011(110)1^ ^

尸'國"//〃八'6 ^0(^61111^ " ' 〃 / !

811X1)16^

^^1100(10118

013^13^ 4.9 — ― ^ 78(116 \ 1^)0(38 10

1(18

! 1(71^011613 0313 5X00^6^1 03(3

| 1 |二國^ —琴"

\

~~—""―

1

1

,」 《0^ 140

1-0313 56|6:|1011 ^

1

0 3 ( 3 50111:6

1

|~|0818 3011^06

冗59

卜 丁 31)15 丁化800《 :"06501'!^11011 [-^111111)6^ 0^ 01)56173110113

29904

1^0(111)67 0『0010111(16

26

卜^9「!131 0318

1-10

|~33110|31|119 1^316 1 !365016111 ! '-30006

1.0031 5/27/061:55

申111113 1-1351 ^101 ;1.331 3(3(115 !" 1^66(^5 1 — 她 」

^0 765

1-7111-18 01 1.3 51 1^1111 "漏1-1051

八|5「01^ 111311~1\ !5 611101^0^ 「0「 I I 化 0 ^ 1 5 1 0 1 1 5 口「015611^ 0 『 I I 化 ! ! ! ^ ! ! ! 0 ^ 1 3 110^^ 丁 1^5

15


0\00^

4: 8"誠^

0^01.^1011 7^66 ^40(1^15 10 ?1^101 8(^0舰011^尺!'^

145

0:5|3|3^ 4.10

19^6(5 | 9(10^ ^1-0138611*1165 | 0601310*15

0601510(1^619^3

56)601 3 ^60151011化^01100: 广1^1101126

1 0

5.0

10.0 0.0

:.

04

丁 1引5 !1131!I'X 15 1156(1卩0「 51山-1「665,

1^601510115 0^0111 110(^2 013551^0311011 3 1 1 ^ 3 5 5 6 5 5 1 1 1 6 1 1 ( 5 0「

丁1化06X1 110^6 15 1116 0 3 * 3 ?;1^1!11011 110^^ 11115 110^^ 013|3|37 4.11 !"。 1^08110 ― 0 ^

0313

12345 10313 5 9 1 ^ 6 ^ 6 0 1 3 9 6 3 1'!'310100

60.0

∼3||。311011

30.0

-1"&51

10.0

-丁!1116 0!' 01-63110(1

5/27/06 1:55?!^

^1111 1。 丄351曰1X0 「 1.351 5131112 ~ 1.166[! 5 10 1^1111 111116 011.331^110 -闩0门0131311011 [!!-1031

丫 65 丫6 5

4, 1 I 5110^5 1116 5011)11^5 0「 1(10 |51*015611105 0『


?1-^3.101^ 1^0(^1魄1177/7 8^8 ^11!0^!~1^

III !1115 6X311113!^, 1 31100312^ 60^0 0『1116

&V:

80(1(110^^!'

5 7 、 3 ^/?/? //^/70775

『0「 1「3111;11〇,30^/0『0『^311^311011 311(1 10 乂0『0!" 5

0

71115 31100311011 15 ^11116 ^ ( ^ ( 「 出 ^ . I I 15卩035115|6 10 1156 ^ ! ^ ^ ! ! ! : ^0|?01110113 31化111011 10 506 II 111056 ^013011:10115 31^601廿16 丁 116【11!!^ 1101^0 111 1110 (^。(^化1^0\^ 61321~3111 15乜化060181011 1)60151011 丁 1X6 110(^6 3!"6 5卜0∼11 111 015^13^ 4.12.

" 。 ( ^ . 丁 116 ;31~0|361"!: 165 0^(110

0:8|3|37 4.12

―啦柳—一.」 陶 0 ^ 0313

… …

0116^ 0 3 ( 3 ^31-131)163 :1门16【31;11^6

";;:"?二 :,::::?」,..,

813^119 1^1116

?1-0601113[! 619^1^1031106 1-6^61

0.2

训551门山65 ;。56 111^11110^06

1^10 2 6

[^门1閥^1 081600^031 8126

5

^11^171^8^ 0^11163

5

&6| 0(51^^09318 8山65 0 501113126 | 3 輝 3 6 論 二

二 : 画 5000

1^01^9 5 3 歸 1 6

5000

^ 0 # 8 6 : ^ :':::::..,. ;

八556551116111

編 副 门则^

0^63^65

1 06^131011 0.25

1 ?5?^1!『0―∼^!)"51^116^11 711716 01 ^333^)1131^16^1

6 6^0!'6 1^0 1 ^63

^3她16

3616011011

1 6 3(^31 ! 31316 1

丫 65 丫 35

1163(^016

[!11116 0 ^ 0 ^ 6 3 ^ 0 0

5/27/06 1:56 ^1^!

I |^0|3!出6 ^6^1111 ^311165 「0!" 311 ^^!^:!^^.八5 0 「651111, 1116 110^65 3「6 5)3111 1151112 ^ (^"―5|^113「6 01'!16!'1011. 丁!16 1|1「65|1013/?-乂^"。『0「 3 5^111 10 1)6 001151(16!'6^ 15 561 31 0.1. 丁 116 5111^-1!'66 56160(1011 13 3006 \^ 00111^1*1113 ^ 6 1 * ^ 6 ^0^11 511106出6 丁 1*66 IV!")10(1 ^1*0(16^ 15 561 10 八88688III6111, 1116八88688016111: 1^638111*6 口「0|3611^ !5 561 10 0 6 ( ^ 0 1 1 , 311^ 3 1)1*0(11: III;""!乂 15 61116!"6丄 1 6

八&6「 「111111;11呂1116 1)00181011 丁1*66 110(16, ^011 0311 0(3611 1116 1^08111(8 \^111^0\^ 5|10∼门111 0|5131^ 4.13.


0X111)1^ 4: 81(11(^1)1^ 0001x10/1 IV1^0^6^

10 1^!~0^101 ^3^0^150

011^尺^

|

献.

一 一 一 "

1*6*1 ! , — 1 — 。 |

3 嘿,

2

0::

一 一 , 一 -

一誉

I I 乂011 1113X1111126 1116 丁「66 ^3116 111 1116 1^6811115∼111110∼,乂011 0811 566 1116 1^6 35 500\^11 10 015^!3^ 4.14. 013)3137 4.14

^011 ^ 6 讽 & 6 66 011 乂011「 301*6611,出6 1111611511^ 0厂【I化00101 0『1116 110(16 1(1(1103105 1116 1*651301156 1*3(0 0卩1116 01151011161*3 !!^^&^ 111出6 0 0 ^ 8 ^ ^3551113出6 11101156 |30!11161~ 0乂6『&0? 00(16 011【1化^15^13^ 1)16 5131151105『0「 11131 110(^0 \^!11 ^6001110 11101^6 ^15!51^ 1:1 015^13^ 4.14 1|化『3「 16&『1185 3 ^ 1 1 6 1330^01111^ 06&51 111101511^ 0? (^(?!"), "!,"、 111(1103(63 1116 (!!^^^! 「65^)01186 &. 丁 1化011510)1161-5 10 11115 110(16 ^0 1101 0\^11 3 1101116 (!"101^15 : 0》,"化^ 0\^11 0116 0「 1∼0 0 ^ 1 1 0;11^3 (^!^!^丁卩:1,2》30^ ^16\ 31-0 23.5 \631~5 0 ^ 0「 乂0111袼61"(八05 〈 23^5\ -

147


?,'^比11、,6 1^06161117^ !!'///? 8/18 &!!^)'/.^

化广1^1-001100! 80111(10118

丁 116 01111111131^6 1 . 1 ^ 011311 !5 5卜0\^11 1)1 1116 8001*6 1^311^11108 0 乂 6 「 1 3 ^

811X111^ /1 ^1100110113

!11 1116 1 0 ^ 16?[-II31101

50011011 0『1110 ^ 6 8 1 1 1 ( 8 ∼ ! , " ^ 。 " , 3 5 5 1 1 0 ^ 1 1 0 1 5 ( 5 1 3 ^ 4 . 1 5 八 .

018^137

4.15八

5^0^6 ^3^1(1^05 0代|1外:怖!)

1 一1 3/ ^= 1 1 1 1 1 1 ︺

0

20

40

80

100

06^116 |―

丁^^I

^ 0 1 1 0 1 1 0 ^ 011 1116 3 ^ 0 \ ^ !11 1116

7^.1.10^7^ |

116X1: 1 0 出 6 13(361 8 6 1 6 0 4

^011^ 6 1 3 ^1-0^-^0^11

11161111 ^ 1 1 1 1 3 1!51 0『0|1311:5 1 0 0 1 1 0 0 5 6 &0111, 3 5 5110\^11 1)1 0 1 5 ^ 1 3 ^ 4 . 1 5 8 .

0;3|3137

4.158

」 110

―''-'……:.^^ 《2.00 一 ^ 1 ^

7

1 50 一

011^|91;76 。仏09^11^61^ ^65^0^156

5^ ~ 25

70181

9^1

八 I I & 1 1 6 1 1 1 ^ 1 ^ 6 V/&^ 0『1113^1118 0113!'15 15 1111^011^1 1116 0 & 1 3 0 ^ 1 1 0 1 1 8 0 1 3 1 0 经 ^ 。 ( ! 。 ∼ . 丁 1^5 0 3 0 1)6 0136116(1 ^

「;—-^化!^!!经!11 1116」1311 3「63, 3X1^ 3616011113 0 ^ 1 3 0 ^ 0 1 1 8

0!5|5|3^ 4 . 1 5 0 . 丁 116 1 ) 3 1 3 0 1 ) 1 1 0 1 1 8 0|;110^ \^111^10\^ 15 5110\^11 111 0 1 3 ^ 1 3 ^ 4 . 1 6 ,

9

^10(10^

3 5 5卜0\^11 I 。


!^0(1^ 10 ?1-^(11^1尺"口。"-化(""!尺/,、'々

0101?!0~ 4: 81111^1/1^ IX'(:^!。"

0:3^13^ 4 . 1 5 0

561601 0^0&. 00111^191^6 1111

~

1 ^0 ~

060)16 '丁∼1闪

~~~側!^丁曰1

013^137 4.16 211

1^:^ 1 ^30^10^^ ! ' "!' VIII '?:,

0 ^ 1 * ^ 0 ^ 1 1 , 。 !'!^!^"'

^ 丁 ∼ 默 》 丁 ^ 丁 …

6651 0^^181(76 丁018 一

86570^

6-^:1 ':、.^、

1

1

二!

561匕【「丁 6631卿8已^)!7∼...

[^"^

865| 1 ^ 1 1 ^

01

^76^3

6已5丁^0円丁 8651 (^訓|61;化义尺… 8651 70131 & 「0(11 ::《':柳」^ ^ , : " : , ' - ' 2 卜』|』,,1。111:

0」^11匕二1^已「60「1丁

0^^01!^6 ^0119

01*1X^117510下;^ ^ ^

0八了^01已

016!' 060116

0 一8已3?一0

尺 印 0 戌 : 0 1 ^!...

0

^6(5011;

1

0^巾…

^『耵 11^1 V 」

0^

||1 0 1 5 ^ 1 ^ 4 . 1 6 , 11161-0 !5 311 X

1 1 1 【 1 化 1 ^ 0 ) 0 0011111111 0

1

1

[

3

1

1

^

0011111111 0 「 1 卜 6 ^ ^ ! ! ^ ! ^ I I ? 丁 0 ^ 丁 卜 1 5 11103115 11131 1110 ^ 6 0 1 1 6 「301^118 15 5卜0化11 011 -

^ 1101*120^1(^1 3 X 1 5 〉

11^1 011311; 5|10\^11 111 「0「 1110 X

311(1 V

6

01111111131^0

11「1 !5 5 1 1 0 ^ 1 1 0 0 1^16 丫 一 3 \ ! 孓 ^ 1 1 ^ 1 111656

1|化8001*0 ^111(111^5

3 X 1 5 , ^ 0 1 1 (;出,^1出『「。「6111:

1(1 0 1 5 | 3 | 3 ^ 4 . 1 3 .

V

III 0 化 1 ^ 0 1 0

III6 乂 - 3 \ ; 5

56^11155

乂011

^011 561201 ^ 1 ^ 6 ^ 1 1 1

丁 0 1113146 1 ) 1 ^ 56|61;1:1011, ^ 0 1 1 1 1 3 ^ 6 1 0

^ 0 1 6 。 。 1 霞 " 0 0 1 6 5 ^ 0 1 1 ^ 111^ 1 0 3 、81*1&616, 3 0 。 5 0 1 6 0 1 1 ^ 6 1*0)6 3 5 5 1 1 0 ^ 1 1 I I I 0 1 5 ( 3 1 3 ^ 4 , 1 1 .

11化 ^31*1313165 011 1116

149


150

、、'!!卜548 &1121'―2

^0^11)1^

^贈:?1-001100!

3^11(11011^/01-

81"!"^^1100110115

018(3(37 4.17 一罾 0^0^^

!^015

下71^ 10 '一"

6已3丁。!^^!!^^^!"

^ 隨 1 。

5

1~ ^、'"圍」1""? "'!0111 8651 0^^1911^6 丁018...

6651 ^^11181^6 1 ^

巳巳丁!^丁。 6

0(1^1005 1)13100

^ ^ ! - ^ 6631 ? 讀

8 6 5 冊

^ ^ 1 0

6651 ^ 66500056 8631 0^111191176 ^ 巳 . ,

8已3冊5?0 ^1^0161-10

8681X0101^0^

^ 0 幽 ! ^ ^65^56 V 01(^1X4 丁 IV已∼0「1 丁

^ 隨 & 0^011110(176 丁016, ^0(^1 01-01113

0^9^8016!'

X 一 ' ― ― ― ―

^11^161':0

0 0 ( 8 ^016

画 8印011: ^ [ ! ^ 01巾…

X ^ ∼ 丁

[^(^[!^

^6^0^1: ^ 7 1 ^ 01田…

V

6 训 11「 丁 口 「IX;

1^1 「

^!10^ III11 !)!!)^『016 853191111161113

6X3111^)16, ^

5 6 0 1 1 1 ^ 1116 1 ^ 0 ) 6 0『1116 乂3「;31316

10 V

3 1 1 ^ 1116 1 ^ 0 1 6 0 ^ 0 2 0 1 1 , 1 : 1 0 X , I ^ 6 1 :


013)3137 4.18 5〔脈14311^1195 0 ^ ^ 8 ^ 5 一

20

40

60

80

100

060)16 '丁^^叫

!

^

^ 0 0 0 ! ' 0 1 1 1 ^ 1 0 0 1 1 5 」 1 3 | 1 , I I I 1 3 ^ 6 1 1116 10(3 4 0 ^ 0 0 「 1卜6 01151011161~5 「 3 1 ) 1 6 3 !^' 1116 |31~6^1016^ ^1-01)31)1111^ 0『「65(301156, 111611 I \^!11 0 3 ^ 1 1 1 1 0 54^0

^0^\\\

0『311 「 6 5 ( 5 0 1 1 3 6 ! " 孓 7 1 1 6 5 6 03(31111-6 「3化5

31*0 0 0 1 1 1 1 1 卬 ' 6 5 5 ^ 6 , 1)1)1 II化乂 31*6 5 1 1 ^ 0 『 0 「 爿 6 0 1 0 1 1 5 1 「 3 【 1 1 1 经 出 6

八 5 13011116^ 0111 63「1;6「, ^ 0 1 1 0311

0513111出6

011111II131 ^ 6

100卜

。3口1 X 1 1 ^ ^

1^63|301136 011311 6^0\\^

5 6 1 6 0 ( 1 1 1 ^ I I & 0 1 1 1 1 卜 6 ^ ! ' 0 ^ ^ 0 ^ 1 1 11161111 5110\乂11 I I I 0 ; 5 卩 1 3 ^ 4 . 1 5 8 .

丁0

3 乂 ! 6 西 0 ^ 1:110 口「山1!118 口「00655 1^15|313^6^1 I?乂 1116 1 ( 6 1 . 2 1 1 0 1 1 1*101, 5 0 1 6 ^ 1 ^ 0 ( 1 6 1 — 1 1 6 1 * 3 1 1 0 1 1 1*101 I I I 1116 1^6511115 ^ ! ! ! ^ ^ , 3 5 5|10\^11 I I I 0 1 5 ^ 1 3 ^ 4 . 1 9 .

0:3|3137 4.19

^10^0^

1:

5(01-61 ^ 加 ^

561601 ^\16

100

丁 3616

(^卩卜^\2^6

丁「66 丁"5

1^3^

00

116卩101 361,6「3163

^

0 1 1 0 0 5 1 1 1 ^ 1116 5611111^5

111

0 | 5 | 3 1 3 ^ 4 . 1 9 15

5)10^11 1)1

013^13^

4.20^

5^


1^^(110(1^6 1^0^11)1^

&1161^!^6 ^1'顯':?1-0011001 8011(110115 ^01~ 81(81)1^^

/^! 1001 !0)13

0;3|3|3^ 4 . 2 0

361601 013(1:卜化「。96 ?1-0(11 ~ 0.0125 ~ ―九?80?^ ^ 0 . 0 1 2 2 0 6 V&??10?

^ 0.011593

0.0100~

0.0075 ~

0.0050 1

0.0025 -

0.00001 5

^1^2

0 1 5 ^ 1 3 ^ 4 . 2 0 5110\^5 1 1 1 6 ^ ! ' 0 1 1 1 门 贈 ^ ! "

15

10

?10(11101.

「0「 1:「665 0 ^ ( ^ 1 ^ 6 1 6 1 1 1 5 1 2 6 5 . 0 ) 1 1116 110「;20111:31 3 X 1 5 ; 廿 1 6

0 『 1 6 3 ^ 6 5 15 5 1 1 0 * 0 111^!031:||1^ 1116 5 1 2 6 0『1116 1 「 2 已 0 。 1116 \^61110&] 3X15, 1116 ^ 6 1 * 3 ^ 6 卩 ! " 。 仏

0『1116 11-66 15

5卜0\^11,

316(3 2 0 ^ 8 6 0 ( 1 0 1 1

丁!16 0310111(111011 0 ^ 1 1 6 3^01~3^ 口「0卩1 !8 !111151!'3^ I I I

80011011

4 . 2 . 5 3 0 ^ III

4 3 5 ,

丁|16 ^!'3(311 5110\^11 111 0 1 5 ( 3 1 3 ^ 4 . 2 0 ^ 6 ^ 1 0 1 : 5 1116 ^ 1 1 1 1 1 1 1 ^ ( ! ^ ^ ^ ^ . 八 1 1 11111131 1「66 XV1111 1 4 1 6 3 ^ 0 5 13 2 1 * 0 ^ 1 1 XV 1(11 1116 1 1 ^ 111! I I ^ ^ 3 1 3 561, 3 1 1 ^ 3 51113~^66 0 『 0 1 1 ^ 7 1 6 3 ^ 6 5 15 5 6 1 6 0 1 6 ^ 115111^ 1116 ^311(1311011 ^ 3 ( 3 5 6 1 . 丁 1 化 5 0 1 0 0 1 6 ^ ^ 6 6 !5 5!10∼11 111

丁 1 化 6 3 . 1 3 1)611 !11^ 111256 0113115 0311 1)6 1^81111:8 " " " ( ^ .

丁1 化 6 ^

!'16^6(1 ^

011 1110 丁31)16 10011『「001【116 1 0 0 1 6 3 ! " 111 1116

15 ^ 「 ( ^ 1|1 0 8 八 8 ^ 3 1 3 5 6 1 1)1 I I 化 ^ 。 ^ ( ; ! :

1)1-11116^!『「0111 I I 化 5 3 ^ 0 0 1 ^ 3 ( 3 5 6 1 ^6 5

丁 1 化 5 八 5 00(^6『0「

乂 4.14.

11 5|10\^5 1116 ^ 3 1 3 ^ ! ! ^ ! ^ ! ! ^ 1116

〇111|3111 4 , 1 15 ^ 1 ^ 6 1 1 111

4.21 一

1 !'6010^^〇11中III; 4 . 1 15 3 口 卜 5 1 1 0 \ ^ 1 1 !11 0 1 3 ^ 1 3 ^ 4 . 2 0 ^


〔!畔!。" 4: 81"!"―

0601x1011 71-00 !^0^615 10 ?1^\0\尺&!扇X。

011^尺!'3殳153

0111|3111 4 . 1

丁 1:3111:

78113:

?1:0^1^亡01;

1

0.000000

0.000000

2

0.000000

0.000000

3

0.000000

0.000000

4

0.004180

0.001449

5

0.008026

0.010144

6

0.008026

0.010144

7

0.012206

0.011593

8

0.012206

0.011593

9

0.012206

0.011593

10

0.012206

0.011593

11

0.012206

0.011593

12

0.012206

0.011593

13

0.012206

0.011593

14

0.012206

0.011593

013^)137 4.21

1 2

3

9

01:00 1?1:1111 電

-

3&118=601113 , 11:66一0111;363 -

1

0

11001)3;

1:1111 ;

4

0111(5111 4 ^ I 3 1 1 ^ 1)10 ^1^1 ^11 I I I ^ ) ^ ) ) ! ^ 80\^011 1 6 ^ 6 3 「0「 1)0(11 1116 、311(1&1;00 1 ) 0 0 1 8 1 0 1 1 丁「(^

4 . 2 0 5卜0∼11131

110(16 5 6 1 0 0 1 5 1卜6 「[^"^-^^(^!

1110 3乂6「3经6 ^ 1 0 ( 1 1 1.630116(1 115 111;1\1|1111111 3 1 1 1 1 1 5 1113^ 1101 3 1 \ ^ 3 ^ 5 1)6 1116 0 3 5 6 ^ 丁 1 化

&0(1

5 ^ 5 0 ^ 011 ^ 3 1 1 ^ 3 1 1 0 1 1 ^ 1 0 ( 1 1 5

0 1 5 ^ 1 3 ^ 4 . 2 2 5 1 1 0 * 5 3 ^01*11^1 11^1; 0 「 1 1 1 6 5001~6 0 0 ^ 6 ~

0111\\


^ 0 0 94008 31(1 ^0 10310335』31(10118 8^01|8 。 卞 — — ( ! ,|3口0111 ― & 卩 八 3 | 5 』 0 ; |3330 ― 腿 5 911; 》119巧!^口。:) 3 ^ 138 6)6(3 ^39(130^^ III 5 口 』 0 3 3 』 3 1 ( 1 36 ^1101 8 6 山 0 1 0 ( 1 0 ( 1 9-1038 0; ^ ; ^ ^ 30 1163 3口03 5!1|丄35140(1834^0 鹏叫0^ ^10X^6 |3^0111 99』1 ^;^;^^!) 8 ^ 0 1 | 81 ^0119339 11; 5 ! 1 1 丄 " 〃 〃 / ^ ? ^ ^ / " ^ 0 1 4 9 ^ 0 ( 1 ^ 9 1 1 3 ; 8 8 ^ 31 I 9 ^ 0 1 4 』 0 1101)1111^9^ 6

^01^91^0(13531|丄 3^140(1^-14014^0 9

/(////^^ ^ ^01^^0(191|;

8; \\ ^98 6)6(3 ^111111&!) 9 ^ 11101;

0

^11(111103 08|8 5; 817179^^25266^6^*0 ^ 0 ^ ^ ^^0(1^94^0 ^111(10(10^(1 51 II 1)116 198 100』】^9;0^11100 81 I 55^99/^^09502*0 ^ 1 ^ 3 9 ^ 011)31181556 5! 1101^/^133(10 116 ^ 0 1 | 5^01|8 ^1^51(3 III ^9^1110111 3|300 8

^0149180(1卩

9

9 ) 0 ^ , I 3^)011 ^0 1113111338 31|丄

^3 ,453^—43 一 453^1

^ 801:1 86&8","? 2 0

, 0 ^ 9 ^ ~ ^

1 3 " : !一 厶

^00

0

03033—3

丁 363;! 一 3

一^31一

~

一3。0『

±0 -咖3

?0(1仪3111《4 0^4 ? 4 ' 、? I丄:;、'3: &

。 一 11:3^1^3一 4 0 ^ 4 》 " 工 ^ 1 "口—11126^一 、^009^0^0、

3^03111 ,!!^隱61"^!13

。 "!1011^9~

^ ^0(1 113111 I 山 ' 丄 ' , & , ) I丄:::3:3

7

III 81 一 ± ^ 9 ^ 一 21

^1110^

誦81一111^3^一 " - 一 "!!3^13~

0

^0(1

"

±0

1

》^11X019 21~^1149^~ 21 3513 ?I 。 "!!3^19~ 《 31100》3"王885111 ^1 ^3 2 7 " ^ ! ^ 1

^00议3111《'?,丄"I ? ('^;【丄《39

^ !""±1133^~ ^^!!^^^^!^ 311011 1X^*1 。 21^X^1^!^

0

" 蘭 《0画 冰"&

3001^

X

" 一

'了

3^一

"

一 1 一 3 ^ 一 口'"忑"';'"

^^3^1

23^八6|33!0


(^叩^

4: 81111(1111^ 060181011 7^66 1 4 0 ^ 8 10 ^6(^101 ^68^0^86

0.11(1 ^18^

155

013^137 4.23

" - , 5

?031:61;101; ? 1:01501(1111165 ;

,-"^ 0^^131011 ? 1:00533109; 101)61 0 — 8 2 3 ? 。 ,。6013100: 2*651?';

181)61 2?^5? ^ ,5^6^601 ^0114: ^63^ ; 1

? 9; 0^3? ~ '; ;2?^81:3?。,; 1

^60X1X181 。 1 ; 31: 0。

^60111101;

0 ^ 3 ? ^ '!'; 2?^825? 。 ?—"^。丄^ 5 ^ 一300

16

^

^^1631)1

#

0十?^163190

3^0言-1; 9

0;

一 3 碰 〉 2 ? ^ 3 ? 屮 2 . 2 7 3 7 1 7 5 - 1 2 1^1611

30;

2?^823? ^ ^31111; —(^(;。!!&! ^ 2; 0 ^ 3 ?

3 ,0,;

611(1;

5113 1)60131011 ?!:0025311113 ;

XII15 00^6 3|10\^5 110^ 360!5|0115 31*6 1X13.6& ^\\\\『6@31^ 10 ^881^111113 30 01)361^31101110 3 131^61 16乂61 ( ^ ^ ! ! 33 1*63(301156 31101 !!^^-^^!!^^). 丁116 36(^51011-11131^1113 151*00055 13 ^13011356(1 111 36011011 4.2.2 30(1 316? 10 0^36011011 4.3.5, 丁 116 36(^51011 131*00653 15 &001 006& 56^111011(8 111 015^13^5 4.22 &1\6 4.23 ^ 10 30乂 & 561 30 06561^31100 13 88313116(1 10 3 00(16 &5 5110^0 I I I 。 ― ― 4.22, 丁卜60, 5356(1 00 1116 0051:6『;0『 131*01)21)111(163 ( ? ^ ^ ^ ) ! 311(1 ? ^^^^) 0『1116 35313116(1 110^6 &11(1 ^0^11 0131『1乂, 6X^01&6卩「0行18 3「6 001111311(6(1 "!!^^! 0601810111 311(1 0601310112. 丁 116 6X1360(6(1 | 3 『 0 厂 " 0 0 0 1 3 1 0 1 1 1 15经!乂60 5^ 1116 5 八 5 81316016111 2?^1123? : ? 一 3 : 6 5 ^ ) 1 * 5 十 ? ^ 1 : 6 5 ? 0 ( 一 : ^ ) , 1^016 11131 1116 011011)61-5 5 30(1 1 31*6 1X0111出6 ^1*31 0011111111 0『化6 ^0^\1 0131「1?^ 1^ 乂011 013531^ & 11*116 !'^?^^^^!" 35 & !^^!^^!!^^!111611 乂00 17131(6 3 ^0^\1 0『55, &1\6 乂00 013831^ 3 00(1-1*6515011(!^!' 33 & 1^51300016『,1)16 口『001 13 - 5 1 . 0 . !"16006, 5? 1158?『6(51*656015出6 0011560)11611063 0『013551(^1113 30 01)561^311011 35 3 ^ ^ ^ ! ! ^ ^ ! " . 丁卜6 31160131^6 0160181011 !5 10 013331^ ^ 0(5561^311011 35 3 000-『6513000|6『,300111113 13 ^1^60 ^ 1116 5 八 5 5(3161716111 ― 3 冊 : ? 一 3 : 6 5 5 ) 1 * 0 十 ? ^ 2 : @ 3 ? 0 0, 丁116 1^0 11111111)61*8 0 &1\6 0 31*6 1116 360011(1 00111010 0『1)16 [?1*0^11 !^^!-^. 7^16 00(^6 5110^3 11131 5001 〉 6?^1108? 1)16111卜6 ^0151011 15 10 355^0 1116 131-861 1^61 0 10 1卜6 01)561^31100 ^ 丁 116 ^3031?16 !11(110&163 11)6 ^60151011 15 0—116^?, ^1110^1 13殳65 1116 乂31116 0 1『1116 01)561^31100 13 33313116(1 10出6 13^61 1^61 0, 0『1 1『1)16 01)861^31100 15 353^116(1 10 1116 131^61 16乂61 1 ^ 811106 01)561^3(10113 11131 &1*6 3551@11601 10 1116 531116 110(16 311 113^6出6 331116 口0516『!0『「1*06361111165, 1116^ 3!'6 &\\ 3&81^11&6 1116 531716 1^&I &1. 1

#

1

#

5

4.4.1

7 6 8 1 1 1 1 9 1^10(161 ^ 6 1 * 1 0 1 - 1 1 1 3 1 1 0 6 ^ 1 1 1 1 3 7 6 3 1 0 3 ( 3

I 1156(1 1116 11-31111113 561『0『36^6101311^ 1^16 110^6 ^611111(10113 311(1 「0516『10『^1*01)81)11&1\6 I 11563 1116 1(13(1011 6&13. 561『0『131*11111113 1110 11*66 1 。 1111(1 1卜6 0(3(111131 1^&. 1^0^ I 116601 30 11X1613611(16111 6&1& 561 00 ^111011 10 6乂3111泔6 1^16 1110(101 ^ 丁 116 1651 (!&561: 15 561 331(10卩0『1(113


?1-6(110(1^6 ^10^11/1^

XVIIII ^ / ! ^ ^11161^136 ^1!10:~ ? ) ^ ^ !

8011(110)15 /。,'

81181)16^ ^4;^11061110115

&556551116111. I XVIII 001111601 1116 !\101101 00111)1:11*18011 1 1 0 0 6 1 0 0 1 6 000181011

0 0 ( ^ 6 3 1 1 ^ !"1111 11.

丁 1化1(68111(8 ^ 1 1 1 ^ 0 ^ 0 「 【 1 化 1 \ 1 0 ^ 6 1 00111(1111*18011 110(56 5 1 1 0 \ ^ 5 廿 化 1 1 0 : 0113115 「0「 3【1 |:|1「66 ^ 8 ( 0 561:5^ 7 1 1 6 5 6 0113115 31'!2 5卜0\^11 1)1 0 1 5 ^ 1 3 ^ 4 . 2 4 ^

0:8^13^ 4.24 5(01^

^3111(11105 0乂6|1財:!^!)

561601 0119^1;

0 0313

^0)8 ~ 1 ^ 1 ^

0

0 15 .01

7國3 , 1力

20

40

0^13

60

[^卜^

40

20

40

I

60

80

V 100

60

80

100

I II

4 一13召 011111=13

1

20

,﹄

\

0

100

八!^

""!

"!

80

060||6 [ ^ 丁 1̶―

丁卜6

I

「!"0「1113111^ 0 ) 1 1 1 6 1110(161 0 0 1116 1051 6^

311(1

1^311011 ^ 3 1 0

4 . 4 . 2

15 ^ 6 1 ^ 51111113!' 1 0 115 |361*101*11131106 011【112

'!"II15 00111031 !5011 51】0∼5 1)131 1116 1 1 1 0 0 ^ 1 15 -

&|3131^119 1 & 6 0 6 0 1 3 1 0 1 1 丁『66

1\10061 1 0 5 0 0 1 * 6

3

X I 1 1 5 5 6 0 ( 1 0 1 1 5 1 1 0 ^ 5 1116 3 ^ 1 1 0 3 1 1 0 1 1 0『1116 ^ 6 0 1 5 1 0 1 1 1「66 1110(161 3 6 ^ 0 ( 5 6 3 !"05(5601 ^0111 丁 116

0 3 ^ 3 5 6 1 10 500^6 3

丁 1化!'600("(!5 111 1116 13^05|3001 ^ 3 1 3 5 0 1 1^0 1101 !13乂6 1116 ^ 3 1 1 1 6 0『1116

1 3 1 ^ 6 1 ^31.181)16 1135 1\^0 16乂6!5, 1.65(30115(2

110

I

^^! !^^]^. -

1152 1)16 1110(^61 1 0 ^ ) ! ^ ! ^ 1116 -


4: 81111^1^ 1)601510)7 7^66 ^0(^618

0^00^

I 1156 1(16 8001*6 110(16 10

I。

^101 ^65^01156

0^

1^13^ 157

I &6^ &VI 10^1111)^(21 110^6『0「

1(16 010^61 &1\6 11131(6 1116

1116 口『05口601 ^ 3 ( 3 3.1X6 3 8001*6 0 0 ^ 6 I I I 1卜6 口『00655 00\^ 3 5 5110\^0 111 015口 13乂 4.25 , 0:5(313 乂 4.25

3 ^ 01:17 0

^ 1003(3 ?3^1111011

^

1^10(161 。01111)3113011

081:13|0门丁『88

0 關

0~ 8001^6

1(1 015口 1&乂 4.25,出6

561111&1 13 1)61113 5001*6(1 13 03016(1 1153?^800118 30(1 115 1^016 15 361: 10 5

8001-6. \^1160 1116 8001*6 00016 15 「110, & 3)3)31163 1116 1)6018100 71*66 0 1 0 ^ 6 1 10 1116 63.1& 561 1153? 8 0 0 1 1 5 ^61161*3(65 ^1*6(110110115 0『1卜6 131*261 1 ^ 6 1 5 30(1 ^ ^ ! ^ 。1*0(73(^111165, &0(1 01*63165 & 5

0 6 \ ^ (^&I& 561 5^ 3^611(11113 1116 131*6^ 1011005 10 1116 |^0515601

561. 0111)3111: 4.2 5卜0\\^ 1116

^31*131)163 111 1116 0 6 ^ 6&1& 861.

0111(3111 4.2

3 0 0 1 6 0^^111: 781:181)163

II磁

?11110X1011

0 ^ 7 0 ^

1)^5?

1)610151011

71:66

01^01&351^1^710^

01^551^10^X10!!

? 126(1101:1011 ^ 0 1 10^3^

1)^X131019

1^600101116X1(16^ 1)60131011 ^ 0 1

?1^1)107

30016

?!151)107

50016

0601310X1: 1163 口

? 101)013111^ 0 ^ 0 1 8 3 3 1 ^ 1 0 ^ 1 0 1 1 & ^ 6 ( ^ 6 ( 1 ?1:0^1^ ^ 0 1

&331:33 01-32&10:訂

" 印

? 1:01)0131111:7 ^01: 1 6 7 6 1 1 0 ^ 1 6 3 ^

" 印

500X16 71: 66

1^1: 6 3?

(:14331^1"丁丽

71166

1X1X^0:"冲

卩一 11631)0

?^0107

71:66

? 1:6(11006(1:

^3^=0

?一 1:631)1

?1^0107

丁 1:66

? 164110^6^:

1163 口 二1

丁 166

11111101:111&1126(1 111^0:

3》

一 励 2 :一

而 !^江丁

7811(3&^6(1: 1:6311=0

?1^1)107

7811(1^6(1:

71:66

^3^=1

歸 6

71:66 一

1:63^

1:1111193

丁 116 131*0^101601 131*0686111(165 31*0 ? — 3 0 ( 1 ? — 7 1 1 6 5 6 31*6 1116 ^05(6!'10^^01)0,1)1111128 01*63(6(1 10 1116 1 「 & 口 『 0 0 6 5 5 , 丁 116乂 31*6 |3311: 0『1卜6 0160151011 11*66 0100161. 丁卜6 8001*6 110(16 ^ 乂 6 5 1卜6 1)1-6(11016(1 ^311165 0『1116 1 3 ^ 6 1 16^613 1)356(1 0 0 ^0^11 0 1 3 X 1 0 1 7 1 1 6 5 6

|31^0^1016^ 乂31065 31*6

5101-6(1 3 5 1116 ^31^13616 5 1 ^ 1 0 6 0 1 5 1 0 0 ^ 0!5口 13乂 4 . 2 7 5*10^5 1116 ^60|116110^ ^^(!-!!)^^!! 0卩1卜6 131-6^101601 131^61 16^6(3 1)2186(1 0 0 。『0111 013X10112&1100^


? ^ 1 0 ^ 1 1 ^

11-/7/7 5/(5

801X11 !01^ /01.

&1161^!156 1\41110:~

811.^11163^ ^^1100110913

013(3137 4.27

I 翁 0 『 0 ^ 5 9 ^ 6"―。3104 5^4^0601 ^!^^^^!:00110611(1611 06(1310门「01『61(11 0

0^5,00 1: ?10!11即化,^001^111 ^ ^10^1

0 28765

0

1110 1119

287&

1119 19904

^6^1^101)0

丁 0131

& 丫 4 . 2 7 5 1 1 5 11131, I I 11(^6!'卩「。厂"1113\1111!2311011^

1 70(11

58785

1 1 19 0111 0『1116 10131 2 9 9 0 4 0115101116「5 31*6

^ 0 5 1 8 1 1 3 1 ^ ^ 3 5 「65口011^16「、 丁 116 5001*6 0 0 ( ^ 6 3 ^ 6 5 311 ^1(6111811^6 ^ ^ 1 0 1 1 0 1 1 0 『 出 6 1 3 ^ 6 1 1 ^ 6 1 5 1 ) 8 8 6 ( 1 011 6「「0『1111111111 !?;^!:^!!. !!" ^ 1 6 ^ 1 0 1 6 ^ ^ 0 1 3 3 ^ | 1 1 ^ 0『11011扁「65|301156 15 ^1~631:61~ 111311 1116 1 3 ^ ! 0 1 6 ^

1^.01^131111^ 0『1-65(301156『0「

3 1 1 ^ 「600「3, 1 1 1 0 1 1 出 6 乂 ^ ^ 匕 ^

8\1^013551^0311011

15 5 ^ 6 1 1 1116

0

『0「 1)131 「0^01"(! I!' 1116 131-6(11016^1 |3「01^11^11^ 01"!'651301156 15 ^ ! " 6 3 ^ ! " 出 3 1 1 1116 131-6(11016^ ^ ^ ^ ! ) ! ! ! ^ 0 卩 11011465口01156『0「 3 5 ^ 0 1 1 「6001^, 111611 1110 ^ 3 1 1 3 ( 3 1 6 8 ^ 0 1 3 3 5 1 1 1 0 3 1 1 0 1 1 「0「 11131 ( ^ ( ^ 「 厶 1 5 ^ ^ 0 1 I I 化 VII1110 1 ^ 丁 11115, 1116 乂 ^ ! " ^ ! ) ^ 5 ^ 0 1 3 5 5 1 1 " ! 0 3 1 1 0 1 1 ! 11^ 1 0 3 ( 6 5 、乂1161|化『3 ^ 0 0 ^ !5 ^ 6 5 1 ^ 1 1 3 1 6 ^ 3 5 3 「651300(16!" 0「 3 0 0 1 1 - 「 6 5 ) 3 0 1 ^ 6 「 「 6 0 0 「 丄 乂 4.28 6 1 ^ 1 — 0 3 5 5 ^ ^ 0 1 1 :31\65 1116

0

5110^8乜131, I I I 1卜6 ^ 3 1 3 5 6 ^ 出 6

V&!'131)16

「0!" 29867 !'600!'(!^, 3 1 1 ^ 1113^65 1116 ^ 1 0 6 1 化 ! " 37 「&:。「(^.

丁 11115, 3 0 0 0 ^ 1 0 5 1 0 1116 ^ 6 0 1 5 1 0 1 1 5 1?356^ 011 2 汀 0 「 111""!;111;231;01,, 0 0 1 ^ 3 7 0 1 1 3 1 0 1 1 1 ^ 5 (「。。。「^^) 3「6

29867

^651^11316(1 35化5口01136「5,

1;115【01116「5 3「6 ^ 6 5 1 3 1 1 2 1 6 ( 1 3 5

116 ^ 6 ( 1 1 1 6 1 1 0 ^ ^ ! ? ! ! ! 1)111 !011 0 『 5 1 ^ 1 ^ 0 1 3 5 5 1 ^ 1 0 3 1 1 0 1 1 -

11011—「65口00^161^

!5 5110\^11 111 0 1 5 1 5 1 3 ^

4.28,

015(3137 4.28

7116 ^950 ^1-0061^0^6

^!.^^^^!^!!^^! 3 ^ 1110(1011『0『1-68111 0 ^

95(10

060:5!0^1 1:尸-『65口1 5 70^1

丁 1化131-0^1-3111

11131

化 ! " 3 ( 6 ( 1 0!5|313^5

4.27

011(1 4 . 2 8 15

1

70(31

29867

0

29867

0

37

29867

37

37 29904

5|10\^11 1)1 0 1 5 ^ 1 3 ^

4.29^


0^006^

4: 81411^1^

1)60181017 7^66 1^0(^18

10 ?^1011^68^09136

091(1尺滅

159

0|8|3137 4.29 9^0^ ^0130&1.; 1 2 ^ 0 X 5 1 0 / 2 1: ?一"65卩2 ^ ?一了65^)0, 0 。 ' ! ? 1 5 0 : ?3^1 ^ &5^0 ; V&丄"6印&!!亡 '! 9 3 1)^0151011 1: ^ 0 X 2 6 ^6015X0**2 1 ^ ^011^ ^01310^ 0 0 2 1)001310^ 0: ?2-0/22; 1227^^2: 015101: 1 ^ ?!'0x11 1^3^&2: ^301 &1022 0 9

9

9

9

1

9

【XIII ;國.^

― ― ― ― ― ― ― ― ― ― ― ^ ― ― ― ― ― ^ 一 一 -

―――――

匸&^16 ^^3^*^^1)^013^0X1詹1101^1:0611^ 1101:0X9 110001 11^16 '磁"漏5 3^52? 08 0?17 !]^12X1101^ ; 10^^ ^^3^ ^ 仏 ; 0(13 1 1 ^ 1 01096

361 6&03 ^ 3001:6^3001:6 ; 01 ^ ^1631)1 口6 ?一"31)0

01:00 1X^9 ( ! & 仁 ^ 化 即 ; 101)16 0 1 * 0 1 ^ 0 1 ^ 8 3 1 ^ 1 0 ^ 1 0 1 1 I 1101)61:6611^ 1101:0^ 110001 ; 亡01:11&仁01 1:丄!;!!!^;.;

0^3 ^!!丄01036 ;

4 . 5 06^0|0||||19 8 869『088誦011 7 『 6 6 0110116誦1。 ?!'6^1101 ! ! 她 10 @61161^1 ^化15

10 11115 56011011 I 0 1 ^ 6 1 0 ^ 3 ^ 0 1 5 1 0 1 1 11*66 1110^01 10 口 巧 ^ " 5

1

6&^0(1

3 3 1116

6 X ^ ) 6 0 ( 6 ^ 30101101 0 ^ 1 0 5 5 10 1卜6 0 0 0 1 ( 3 3 0 》 八 0 311(0 1115111^1106 0 0 0 1 ^ 3 0 ^ 1000『5 103365 1*0511111113『『001 5110)1 6 ^ 6 0 1 5 3 3 3001(1011(5 0『1卜6&5^八II 311(0 1113111 31106 00111^311^ 13 0 ^ 6 0 111(01*6816(1 I I I 1

6X^60(6(1 1。88介叫"6""

1)1*6(110(1113

1116

0『1(3 01131011161^ 0『1?^05136015^ 丁 116 1113111*31106 0001|5311^ 0 3 0 3611^16

131-6111111111 ?0「 3 。0510016『&000^111^ 10 1卜6 ^6x0X^(1

1055^6^146^0^

1085^6^146910^

15 111638111 6^ 3 5 1

1116 ! ) ^ ! ! ! ^ ] ' 0『1088&8 |56!" 0&1- 乂6&!", ^1161*6 03!" 乂68『15 (^行!^^ & 8 也 6 ^ ^ ^ ^ ! ! 0 ^ 11)6 1115111^1106 |30110^ 111111(1^116(1 \ ^ 1116 !!"!!!!^! 0 ^ 031*3 1115111*6(1

1116 口011。乂. 1『1卜6 1113111^1106 ^30110^ 1135 1)660

1

111

01『0『『00『01001115 81\6 0 0 6 03「 13 00^61*6(1, 1116111116 1111111136『0『031^ ^631*3 !5 0,3333, 1『11161*6 15 15 1/ (々门之):3, 1^ ^111*1113 111656『011「

0 0 6 1053 ^111*1(13出15『00『01001115, 11160 1116 1088^6^146^

1110111115, 1 ^ 0 031*3 3『6 00^61-6(1 "!!^^! 出6 53016 口01!0乂, 11160 1)16 0001136『0^ 03『乂6&!"5 13 0 . 6 6 6 6 ; &1\6 1

1『1(161-6 15 0 0 6 1055, 1116 1058^6^146?10^ 15 1.5.

丁 116 131*861 ^31*131)16 !0 11113 1\\0^6\ 18 1卜6 1055^6(^11611^ 13 00111;11110115, 1卜6 11*66 00115(1*11016(1 13

3

311(1 15 00111111110113. ^11011 1116 131*^61 ^ ^ & ^

6551011 1^6^八^?/。"

^6, 111(6 & ^60151011

1『66, ^3111110113 1116 1*6001X15 (。!^&!!^!^) 1010 56^111611(5 0 『 " 。 ( ^ . 丁 116 ^6^65310!1 00051315 0『1116 1&0/\^\\\\^ ^60181011 1^66動(^!, 1^16 6X^60(6^ 00(16 111 & ^6^6881017

1^66鹏10

^0^X0^^01)01)111(168

^66觸

&1*6 0^101112(6(1『0『030)1 160/110(^6 I I I 3

^146 。『1716011 0犷1)16 1 3 ^ 6 1 ^31*13516 13 03101113(6(1『0『63011 1116 01117601 6X3111 口 16, 1116 ^ 6 6 03101113(65 1116 6X13601:6(1

1X6^116110^ 0『105863『0「 63011 160/?10^6 0 『 5 6 ^ 6 1 1 ^ 711656

丁116 |3^00655 ^0\^『0『1110(1611113 「!5^ 15 8110^11 !11

乂 430^

11163113 31*6 0 3 1 0 1 1 1 3 ^ 6*001 1(16


015(313乂 4.30

^

^15^:

0

^

^

0

̶

^

060151011 丁「66

0313

~

~

1^01^61 00111 !]31'!3011

丁 1化1?1~0|361^ 301111125 0 ^ 1116 1)0^181011 丁 1*66 110(^6 31^6 5!10∼11 !11 0 1 5 ( 3 1 ^ 4 . 3 1

1|1 11115 6 X 3 1 1 1 1 1 化 1 1 1 0 ^ 6 1 1 1 1 8

5 6 1 !5 |53111110116^ 311011 111&I 60^0 0『1116 「 6 0 0 1 ^ 5

113111!115 3 0 7 0 & ! " 乂311(^31100, 3 1 1 ^ 1 0 ^ 0 『 0 「

7 1 1 6 8616011011 11161110^ 1 1 5 6 ^

0

5

3110001001

^01~ 11115 ^3(11110111113

18 511111310 「311厶0111 53111(3110呂.

0:3|3|37 4.31 9^0^

791116

:^0^6.10 巳 6 3

03!9

?'13^163 1^119^011^6

^ ^ ^ ^ ^ ^ ^ ^

朋 ^ 騸 ?;116 !'100 |~|31911111031108 1-6^61

0.2 038…5331X11

川。56 1 , 1 0 ^ 6 1,3乂1謂^!日'謝^!

1^0 2 6

^ 圖 1 1 1 ^ 1 03(600[!03! 5126

5 5

: 咖 ^ 6 「 0(1^165 ! + " " 議 1 3 6 「 0 ^ 【 「 0 9 3 1 6 ^^185

5 0

1-301118126 國 辈 5 ^ 3 她 5000 " ' 圖 6 33111016

5000 ^336381716111

1-^111111136!'0(1.63^83

1 ^61-396 5^312 &「0「 0.25

「:曰01^6『「0山^1^51(716111 1-111108 01 ^355 ^11^8101601 !~!|1101115

丫6 5 86^01-6 卜10

! 「:|^111&6「 01*1,15 30!11她1511^16111

1

^65 罾罾

卜:&16

6618^11011

^

,83

卜匕331^3 「131516 :5 69 018111

^ 7 1 1 6 1113111 ^ 1 ! ^ 6 ^ 1 1 0 6 1 3 6 1 ^ 6 0 1 1 1^665化化】1^111^17

^ ^

6

1

53 1 1

^ 忭

6

6

5

"; 卜00111111110115 1 ^ 6 1 5 !5 1 1 1 出 6 1

1651: 5 ^ 1 5 1 1 0 5 1 1 5 6 ^ ?0^ 6^^11131111^【1化5^1115 311(1 1 卜 6 03|01|131!011 0尸1116化0「111 0『1116 11-665. 「01 1116 -

00111111110115 13「^^, 1116 1 0 5 1 5 ( 3 1 1 5 ^ 0 卜 3 5

311卩―爿^忭化1111011.

丁 1 化 1 1 1 0 ^ 6 1 0556551110111 13 ^ 0 1 1 6 ^

1151113


0101)1^ 4: 8111^1111^ 060^10(1 丁,'66 ^10(16(5 10 1^^101 116^0)1^

1116已「「0「

5111115

^ 1*1*0!". 4.3.8

0^

? ^ ^ ^ ^ . 丁1化81)111(111^

01*1*61-1011 [^。(^;,"!〉'

丁 1 化 ^ 8 8 6 8 5 1 1 1 0 1 1 1 1^1038111*0 ^ ^ ^ ^ ! ^

111(01^31

4 . 3 . 1 ^4 ^!50115565 1116

!5 ? 1^05^

^111011 15 1 1 化 ^ 6 1 0 1 1 1 1 「01

-

「0「 3556551116111 15 5 6 1 1 0

1)1*01)1^ 01'!161'!011 「0「 ( ^ & 山 ^ 山 ^ 出 6

011^11^1^

0 1 - 3 ^ 0 8(111;! 1*0

\\0011

0「3

80011011

0)150115565 1 1 八 \ ^ 「 3 ^ ; 5^ 1131*6已「「0!" 15 0310111.116^^

015^13^5

4.32八1111-011811 4 . 3 2 0

5|10\^ 1110 !1*00 1 1 1 0 ^ 1 ^ 6 ^ 6 1 0 ( 3 0 ^ 1 ^

0 — 1 3 乂 4.32众

"&, ,1|1∼1,。

0:8^13^ 4 . 3 2 8

1116 0 0 0 1 8 1 0 1 1 7 1 6 6

1 1 0 ^

161


015^)137 4 . 3 2 0

, 匕 0 亡 丄 7

= 1*10103

^751:050 I

,""漏"0 ", 8 土0 1^61

7^*1X1^^9 7 , 1 " " 丄 0 0 0-21 0-20 1621 3 83

0 ―】

91*^1?^^^

^01.111119

VI 1丄。"丄011

?!:&!^!^^

^0113醫【丄0「

7 1 1 6 乂 3 「 1 3 6 1 6 5 ^0011!11@ 1116 110(^05 0 ! " 1 卜 6 11*66 3!'6 ?"!?11X^10 ^11II1111^6!' 0 「 |3卜0「 乂;。!。"。!,;),

01^50

^0!'6(^11: 5 0 0 ! ' 6 3 1 1 ^ 八 0 5 ( 八 ^ ) '

0 1 5 ^ 1 3 ^ 4.33

5 1 1 0 ^ 5 卜 0 ∼ I I 化 1 1 1 0 ( ^ 1 8110066(15 111 !'3111(1113出 V&!"10115 1 * 1 3 ^ 吕 謂 口 5 ^613 6

10 0 1651 ^ 3 1 3 5 6 1 .

0151313^ 4.33

!5 ^61161-31601 6乂

1^10(161 001111131^18011

1103已

3(3卩"6《


0:3^!3^ 4.33

5

0

0 5 0 0 0 0 20

40

130

60

100

061:116

^ 丁巳; :丁

0^1^

丁 112 11112 0 1 1116 1^0811118

1 1 1 0 0 ^ 5 1 1 0 ^ 1 1 I I I 0 | 5 ^ 1 3 ^ 4.33

( ! ^ ! ^ ! ! ! ^ , ^111011 1 1 1 6 3 5 1 1 ^ 5 7

7116 ^ 0 ) 1 0 5

11^10^『0「 6 3 ( ^ 1 ^ ! ! ^ .

丁1化^ " ? ' 3

0 1 0 3 5 1 1 ^ ^ 3 5 3 00111111110115 乂3!";

I

I

1

0

5

5

3「6 5 1 1 0 ^ 1 1 011【1化

^ 6 ^ 1 1 6 1 1 0 ^ 111 1116 I I ^ 6 0 1 1 6 3 0 ^ 111050

1 1 1 6 ^ 6 1 1 1 0 3 1 3 X 1 5 5 ^ 1 0 ^ 5 0 1 6 0 1 6 3 1 1 0 「 30(11^1 !0^

XVIII! 1116 1 0 ^ 6 5 1 1 0 3 5 6 5 1(1 1116 1611(11 1

15

0 ! " 1116 ^1-6(^1016(1 !0531^110^

00111(31116(1『「0111 0 ^ 6 ^ ^

1101'!2011131 3乂!5, XV 1(11 111056 0115101116「5 ^1(11

//'^

15 8001*6 1^311^11138 0 ^ 0 1 * 1 3 ^ : (^乙!^?!之^.

/0.V.V^^^(/^'/;^'V^, 15 " 化 1 3 「 ^ 1 311(5

1

5 1 1 0 ^ 5 【 1 化 1 1 1 0 ^ 6 1 5 ^ ! ) ! ! ! ^ 1 0 1*316 1116 「 ; 5 殳 3 5 5 0 0 1 3 ( 6 ( 1 7

01151017161- ^!"01屮5, 1 「 出 6 1110^61 1*316^ 1112 「;5^ ^ 0 1 1 1 ) 5 0 0 1 ^ 6 0 1 1 ^ ^ 1(1611 1116 10(3 ^ 6 0 1 1 6 ( ^ ! ! ^

5\\0\\\6

1 1 ^ 6 1116 !!!^1651 3011131 1116311

1。38介&^"(:》丁

1055^6^14611^

。 ― 1 巧 4 . 3 3 5 卜 0 ∼ 1 1 1 3 1 1)16 [!10^01 13 0 0 1 1 ^ 0 1 ! ) ' 「 ^ 。 8 出

3000^!!1^【

7 1 1 6 ^ " ?' 1

0

3

1 8

11化3^66 0 ^

「^.

1 化 1 1 6 X 1 1^60!16 ( ^ ^ ! ^ !)

^ 出 1 〉 5!101|13

\\^\6 8 1 0 \ \ ^ 「 1116311 6

!11(^^(^11315


164

^0(^61111^ \ ^ 8/18 ^11161-^156 1^1116?': ?!'00(100180114(10118^

^60/10^6

311111111317 0 1 1 1 1 6

4.5.1

8118111688

^^110(1110^

1169^881011 7『66 1\11036誦I。 。 1 * 6 ^ 1

八1111011^1 1110 ^ 8 ^ 6 5 5 1 0 0 打 6 6 5110\^11 111 0!5口I&乂5 4.32八出『011^1 4 . 3 2 0 13 ^^\0^^ 31)111113(6^

1)1311 &001

11 01X^1(565 3 1*63115110 111115(1*3(1011 0 ^ 1 1 0 ^ 11113 1116(110^ 0311 1)6 1150(1 1 0 1^60111^ 1116

丁 116 ^ 6 ^ 6 5 5 1 0 0 11*66 1710(161 ^6^610)36(1 111 ^ 1 5 560(1011 1138 1)6611 1136(1 10 ^011|5 1116 !"6001^5 ( ! ^ 11115 0356, 1116 00510016!"5〉 10 1116 6^

561 1010 10出5』0101 ^^011|33~~006『0『63011 0^1116 10

110(165 0『

16111111131 00(168 0 0 1116 1 1 * 6 6 ~ 3 5 5 1 1 0 ^ 0 111 0!5口1&乂 4.34 15610^^八5 口『6乂!01151乂 016011006(1, 111656 2 「 0 叩 5 31*6 03116(1 1^16 1 0 ^ 6 3 0『161*0111131 0 0 ^ 6 5 0『1116 XV&&. 7\\6 ^ 6 6 1110(161 &130『6|30115 1*16 0 1 6 3 0 0卩 1116 131-^61: ^31-131)16 & 『 6 3 0 1 1 ^0X1^ 0^101118(6(1 6*001 1116 11*310108 11163115 &『6 1156(1 3 5 ^60110110115 0『1116 I & ! ' ^ ^31*131)16 1053

561, & 0 3 111 11)13 6X31111)16, 1)1656

11600^《0『6&0^1 0 ^ 1116 1 0 8『00口5,

811106 1116 1116&II 0『1卜6 1^1^61 V&!"13616『0『630)1 31*011^ 0311 1)6 1136(1『0『[^^(^(^。!!, I 11^6『6&1*1*6(1 10 II: 3 5 2X^)60(6^

111 0 1 5 1 ) 1 4 . 3 4 ^ 丁卜6 6X^)60(6(1 1033 ^1*601116110^ 0尸& ^0\1^ 1*6116015出6

1088^6^146^

16乂61 0『1*13^ 833001^16(1 0 『 I V

1116

丁^

16

8『00口5 31*6 3110^11 111 015口1^ 4.34 111 ^001*6&81113 01^6「

1*13^ 16乂615, 015|513^ 4.34 &150 8110^8 1:116 ^ 0 1 1 1 0 3 0 ^ 1116 011^61*601: 1*13^ 8『011135, 111&I !5,出61『

(^&!^。^!^^。? ^\\\\ 1 - 6 8 3 ^ 10 ^ ^ 1 0 1 1 3 乂!01&11003, 01*6(111; 500『6,

018|3|3^ 4.34

^ & .

!!"。川65 0, 005|011161^5∼1出011行6「6门1 「15^ 1*311^3什01X1 1116 1=169!'6331011 3

丁「66 1110^81

1.63(10

140(16 10

1

9

14

2

2

3

5

4

10

5 6

(^化

11/16311 0^ 018 丁3【081 ( ^ ^ 饰 ! ! 1058 (『。!!" ^巧)

,叩&quo