Support vector machines through financial time series forecasting

Support vector machines (SVMs) are promising methods for the prediction of the financial time-series because they use a risk function, consisting of an empirical error and a regularized term, which is derived from the structural risk minimization principle. This study applies SVM for predicting the stock price index. In addition, this study examines the feasibility of the applying SVM in financial forecasting by comparing it with the back-propagation neural networks and case-based reasoning. The experimental results show that SVM provides a promising alternative to the stock market prediction.