Hybrid fuzzy-EC systems in simulation & selection of stock experts

The opinions of different experts in the stock markets often divert with each other. Basing on these various opinions and assumptions, the prediction models and results are also very diverse. They will reach their conclusion from different angles. For example, one may say that Hong Kong´s market is influenced heavily by its own past movements, while others may argue that US may also has a strong influence. The previous studies like [S.I. Ao (2003)], [L.T. He (2001)] supports the later one. There is strong correlation between the US market and the Asian markets in the long run. The VAR analysis shows that the US indices lead the Asian ones. But, in my previous study [S.I. Ao (2003)], such correlation is found to be time-dependent and affects the performance of using the historical US data to predict the Asian markets by neural network. Here, a simplified automated system is outlined to overcome this difficulty by employing the evolutionary computation to simulate the markets interactive dynamics. The aim is to supplement the previous studies, which have focused more or less solely on the local stock market´s historical data, with additional information from other leading markets´ movements. Fuzzy logic is used to generate rules for the prediction. The EC component of the hybrid system is to select the rules most suitable in each situation