SVM application of financial time series forecasting using empirical technical indicators

Support vector machines (SVMs) are promising methods of pattern recognition in financial time series because they use a risk function consisting of the empirical error and a regularized term which is derived from the structural risk minimization principle. This study applies SVM to pattern recognition in the financial engineering domain. Compared with present machine learning methods in financial forecasting, this study does not simply work on the original time data series, but some interesting and empirical technical indicators. The experimental results show that transforming the input data space of SVM can bring good performance in finance engineering.