A novel nonlinear combination model based on Support Vector Machine for stock market prediction

Abstract

Stock market predictions comprise challenging applications of modern time series forecasting and are essential to the success of many businesses and financial institutions. In this paper, a novel nonlinear combination model is presented for stock market forecasting, which based on Support Vector Machine (SVM) regression combining the linear regression of traditional statistical model with the nonlinear regression of Neural Network (NN) model. Firstly, using different linear regression model to extract linear characteristic of stock market system. Secondly, using different NN algorithms to extract nonlinear characteristic of stock market system. Finally, the SVM regression is used for the nonlinear combination forecasting model of Shanghai Stock Exchange index. Empirical results obtained reveal that the prediction by using the nonlinear combination model is generally better than those obtained using other models presented in this study in terms of the same evaluation measurements. Those results show that that the proposed nonlinear modeling technique is a very promising approach to financial time series forecasting.