Improved Binary Tree Support Vector Machines for Multi-class Classification

被引：0

作者：

Pan, Yuqi ^{[1
]}

Zheng, Yanwei ^{[1
]}

机构：

[1] Univ Jinan, Sch Informat Sci & Engn, Jinan 250022, Peoples R China

来源：

2011 INTERNATIONAL CONFERENCE ON COMPUTATIONAL SCIENCE AND APPLICATIONS | 2011年

关键词：

Support Vector Machine; Binary Tree; Muti-class Classification;

D O I：

暂无

中图分类号：

TP [自动化技术、计算机技术];

学科分类号：

0812 ;

摘要：

The binary tree SVMs(Support Vector Machines) algorithm for multi-class classification can solve the unclassifiable regions that exist in conventional multi-class SVMs. At present, the most of binary tree SVMs algorithm create a skew binary tree, so the sample training time and the testing speed is slow. In this paper, an improved binary tree SVMs algorithm for multi-class classification is proposed. This algorithm can create a symmetric binary tree and increase training and testing speed. The results of experiment have proved that this algorithm is feasible and effective.

引用

页码：111 / 116

页数：6