|
|
|
| A Method for Optimizing the Combinational Kernel of Support Vector Machine Classifier |
| YANG Xu,YANG Xin,XIONG Huilin |
(Institute of Image Processing & Pattern Recognition, Shanghai Jiaotong University, Shanghai 200240, China)
|
|
|
|
Abstract To solve the problem of model selection for support vector machine(SVM) classifier, a featurespacebased class separability measure(FCSM) was proposed. With this measure, the combination coefficients of multiple Gaussian functions were optimized. Compared with the kernel matrix evaluation measure (FSM), the new measure has fewer limitations in the application of kernel optimization, and has better theoretical guarantees. The experimental results show that the proposed algorithm outperforms the crossvalidation method, the radius margin bound method and the FSM based method, and moreover, it achieves better performance on SVM classifier with the optimal kernel selected from a wider range of function set.
|
|
Received: 13 October 2009
Published: 31 August 2010
|
|
|
|
| No related articles found! |
|
|
|
|
