Abstract: With the opening for centralized market, numerous investors have applied extensively the technical analysis method to predict stock prices. The foundation of technical analysis is based on the assumption of self-circulation characteristic within stock market. In a word, investors profit by concluding technical mastery derived from past information. Support Vector Machines (SVMs) are a promising tool for doing classification and regression. This paper uses the regression function of SVMs to predict stock prices. We gather past data, calculate technical indicators, and using SVMs to train and to predict stock prices. By applying the proposed method in this study to Taiwan stock market, Taiwan`s stock market appears not that unstable and six well-known leading stocks all have the prediction accuracy of not less than 55%.
Rong-Chang Chen and Chao-Ming Lin , 2005. Forecasting Stock Prices with Technical Indicators Based on Support Vector Machines . Asian Journal of Information Technology, 4: 14-21.