Stock Exchange Forecasting with the Help of Neural Networks

概要

論文の詳細を見る
For the human brain there are a lot of problems that seem to present no difficulty at all, still they are problems of high complexity when handed to a computer. Neural Networks(NN)represent one of man's approaches of simulating the process of human thinking, with the goal of constructing computers that can resolve such problems as easy as humans can. In a certain degree, NN simulate the in reality existing connections between the living neurons in a human brain, and the synapses between them. But NN can go further than just trying to simulate the human mind : they can solve problems that present difficulty for man as well. One of the possible application fields is economical process forecasting. This study considers the Stock Exchange(SE) Simulation possibility. The study is still in it's early stages. Several approaches have been considered, out of which : temporal sequence processing, filter coefficients computation, the game approach, genetic algorithms, intelligent interpretation of linediagrams, explanation facility of NN, OCR approach, parallel processing. The gain that this study offers is, next to opening a number of search paths for further research(ers?), also a theoretical and methodical problem analisation. This article proposes also an original solution, with it's implementation scheme.
1996-09-04