A New Method for Optimizing The Combinational Kernels

摘要：

Tbe optimal kernel selection is a critical problem for the kernel-based learning algorithm. In order to obtain good results, the kernel function must be chosen in a data-dependent manner. To this end, we propose a new feature space based class separability measure to evaluate the conformation of kernels to the data. The optimal combination coefficients of multiple Gaussian functions are obtained by optimizing this measure. Experimental results show that our algorithm outperforms the cross-validation method and the radius margin bound method, and moreover, can further improve the performances of SVM classifiers.

关键词： kernel method combinational kernels kernel optimization pattern recognition

作者: XUE TIAN XU YANG

作者单位: Institute of Electrical and Mechanical Engineering, Jiaxing University, Jiaxing 314001, China Institute of Image Processing & Pattern Recognition, Shanghai Jiaotong University, Shanghai 200240,

会议类型: 国际会议

会议名称: The 2010 International Conference on Computer Application and System Modeling(2010计算机应用与系统建模国际会议 ICCASM 2010)

会议地点: 太原

会议语种:英文

页码: 654-658

在线出版日期: 2010-10-22（万方平台首次上网日期，不代表论文的发表时间）