DealListopad Marzec 2010 Deal 2010
Forecasting financial markets using neural networks
This article examines and describes the use of neural networks as a forecasting tool. It is a survey based on the application that predicts stock market prices using artificial intelligence and its ability to discover patterns in nonlinear and chaotic systems.
hile not only briefly discussing neural networks theory, this study determines that model accuracy is difficult to obtain and can take months or years of investigation. It clarif ies that neural ne t work s a r e not a lw ay s p er fe c t i n their predictions. Nevertheless, Edward Gately in his book Neural Networks for Financial Forecasting describes the development of a neu ra l net work t hat achieved a 93.3 percent probability of predicting a market rise, and an 88.07 percent probability of predicting a market drop in the S&P500.
de sig ne d i n a m a n ner of a biolog ical neurons. It is well known that all inputs to the cell body of the neuron arrive along dendrites. These can also act as outputs interconnecting inter-neurons.
napses that connect it to the dendrite of another neuron. When the electrical input to a synapse reaches a threshold, it will pass the signal through to the dendrite to which it is connected.
Mathematically, the dendriteâ€™s function can be approximated as a summation. Axons, on the other hand, are found only on output cells. The axon has an electrical potential. If excited past, it will transmit an electr ical signal. A xons terminate at sy-
A neural network may be also considered as a data processing technique that maps or relates some type of input stream of information to an output stream of data. For example, the input may be in the form of a two dimensional image with missing
Figure 1. Neural network structure
What is the most important, this thesis does not show a successful neural network , but proves that artificial intelligence can be used as a forecasting tool, describing the idea. According to E. Michael Azoff, a neural net work is a computer pro gram that can recognize patterns in data, learn and make forecasts of future patterns. The name â€œneural net workâ€? der ives from the neural str ucture in the brain, and was so termed by the biologists and physiologists who attempted to stimulate and model the neurons. THEORY The str ucture of a neural net work is
Source: Authors realisation
X numer magazynu Deal wydawanego przez SKN Inwestor działające przy Wydziale Ekonomiczno- Socjologicznym Uniwersytetu Łódzkiego.