Neuro-genetic system for stock index prediction

This goal of the paper is introduction and experimental evaluation of neuro-genetic system for short-term stock index prediction. The system works according to the following scheme: first, a pool of input variables are defined through technical data analysis. Then GA is applied to find an optimal set of input variables for a one day prediction. Due to the high volatility of mutual relations between input variables, a particular set of inputs found by the GA is valid only for a short period of time and a new set of inputs is calculated every 5 trading days. The data is gathered from the German Stock Exchange (being the target market) and two other markets (Tokyo Stock Exchange and New York Stock Exchange) together with EUR/USD and USD/JPY exchange rates. The method of selecting input variables works efficiently. Variables which are no longer useful are exchanged with the new ones. On the other hand some, particularly useful, variables are consequently utilized by the GA in subsequent independent prediction periods. The proposed system works well in cases of both upward or downward trends. The effectiveness of the system is compared with the results of four other models of stock market trading.