Journals Proceedings

Abstract

Stock directional classification technique proposed in this paper is to determine the return direction of stock market. Some previous works used a support vector machine to predict the return direction of the stock market index based on first order difference, stock return, as inputs. Our paper will use higher order differences and higher order lags as inputs for a support vector machine. By using higher order differences and higher order lags of time series data, the experimental results show the improvement of the prediction accuracy with respect to the number of lags and the number of difference orders. In addition, results of this technique show better daily directional prediction accuracy comparing with the other models.