Fast Training of Support Vector Machines Using Top-down Kernel Clustering

摘要：

How to deal with the very large database in decision-making applications is a very important issue,which some-times can be addressed using SVMs.This paper presents anew sample reduction algorithm as a sampling preprocess-ing for SVM training to improve the scalability.We developa novel top-down kernel clustering approach which tends tofast produce balanced clusters of similar sizes in the kernelspace.Owing to this kernel clustering step,the proposed al-gorithm proves efficient and effective for reducing trainingsamples for nonlinear SVMs.Experimental results on fourUCI real data benchmarks show that,with very short sam-pling time,the proposed sample reduction algorithm dra-matically accelerates SVM training while maintaining hightest accuracy.

作者: Xiao-Zhang Liu Hui-Zhen Qiu

作者单位: Normal School Heyuan Polytechnic Heyuan,Guangdong 517000,China School of Management Guangdong University of Business Studies Guangzhou 510320,China

会议类型: 国际会议

会议名称: 2008 3rd International Conference on Intelligent System and Knowledge Engineering(第三届智能系统与知识工程国际会议)(ISKE 2008)

会议地点: 厦门

会议语种:英文

页码: 968-971

在线出版日期: 2008-11-17（万方平台首次上网日期，不代表论文的发表时间）