A HYBRID FORECASTING MODEL FOR STOCK MARKET PREDICTION

Abstract

Stock market predictions have been studied by academics and practitioners. In this paper, a hybrid model is proposed to predict the stock market movement. We have combined the independent component analysis (ICA) and kernel methods. ICA is used to select the important indicators. After determining the inputs, kernel methods are employed to predict the direction of the stock market. We have used the Dow-Jones, Nasdaq and S&P500 indices for experiments. Technical indicators of the indices are used as input variables for the proposed model. According to the analysis of the experimental results, kernel methods are capable of producing satisfactory forecasting accuracies and gain rates for Dow-Jones, Nasdaq and S&P 500 indices. The trading experiment shows that the kernel methods obtain higher rate of returns than the other investment strategies.