Stock exchange index prediction based on wavelet-based adaptive support vector regression algorithm

In the paper, wavelet-based adaptive support vector machine is applied to stock exchange index prediction, and Morlet wavelet function can be used as kernel function of adaptive support vector regression model. In the study, the proposed wavelet-based adaptive support vector regression models trained by the training sample sets with 2~6-dimensional input vector respectively are used to show the superiority of the wavelet-based adaptive support vector regression model compared with adaptive support vector regression model. The comparison of the prediction effects for stock exchange index between WASVR and ASVR trained by 2~6-dimensional input vector respectively show that the proposed wavelet-based adaptive support vector regression algorithm is better than adaptive support vector regression algorithm in the stock exchange index prediction.