Optimization of Technical Rules on the Basis of Intelligent Hybrid Systems

The stock market is an example of a complex system where process of decision-making involves great difficulties, since numerous inherent factors and a high level of noisiness influence the stock indexes. Technical analysis (TA) is one of the most popular methods widely used by stock traders (Dempster and Jones, 2001). TA proposes a set of numerous investment rules to the investor. The use of individual trading rules is not effective from the investor’s point of view (Dempster and Jones, 2000) while application of numerous rules is complicated by optimization of separate rules and the problem of their effective combination. The optimization approaches based on evolutionary computations and artificial neural networks (ANN), have gained wide popularity in problems whose solution space is so complex and large that it is impossible to use the traditional methods of optimization. Each individual technique has its advantages and disadvantages. For example, Giles and Lawrence (2001) showed that ANN could be successfully used in short-term series prediction. The application of ANN is especially effective when the link between the influencing and the dependent variables is nonlinear and very noisy, which is typical of a stock market. The main drawback in the ANN application, however, is that they do not give us any understanding of the underlying process and prevent us from obtaining a specific collection of rules, therefore decision-making relying solely on ANN functioning is not desirable. On the other hand, the application of genetic programming (GP) eliminates those problems and allows creating more complex rules out of simple ones, but it requires great computational effort. The use of hybrid systems could help to avoid the weaknesses inherent in each methodology, while capitalising on their individual strengths. In this paper the author proposes his own approach to optimization of investment rules on a stock market combining GP, ANN and TA.