Applying two-stage XCS model on global overnight effect local stock prediction

Abstract

This study applied an integrated artificial intelligence method, extend learning classifier system (XCS), to predict the stock trend fluctuation considering the global overnight effect. However, some researchers have already indicated that XCS model that is applied successfully to form a forecast model in local market. Based on those prediction models, we put more effort to focus on the financial phenomenon, overnight effect between each two global markets, and we developed a two-stage XCS model to forecast the local stock market. In the experiments, DJi and Twi are chosen as referent and predicted markets respectively, and the model is trained by their historical data. For its accuracy verified, the model is tested by recently data. Finally, we have concluded that the proposed model successfully simulates the phenomenon, and the high ratio of correctness is definitely figured out.