Wavelet Kernel Twin Support Vector Machine
To enhance model’s ability of reflecting data distribution details and further improve speed of data training, a new wavelet kernel is introduced. It is orthonormal approximately and can save more data distribution details. Based on this kernel, a wavelet twin support vector machine (WTWSVM) and a wavelet least square twin support vector machine (WLSTSVM) are presented respectively. The theoretical analyses and experiment results show WTWSVM and WLSTSVM have better performance and faster speed than those in the existing works.
KeywordsTwin support vector machine Wavelet kernel Least square Nonlinear
This work was supported in part by the National Natural Science Foundation of China under Grants (61472307, 51405387), the Key Research Project of Shanxi Province (2018GY-018) and the Foundation of Education Department of Shaanxi Province (17JK0713).