首页 | 本学科首页   官方微博 | 高级检索  
     检索      

树形结构SVMs多类分类的研究
引用本文:王春歆,李连,张玉叶.树形结构SVMs多类分类的研究[J].海军航空工程学院学报,2005,20(2):254-256.
作者姓名:王春歆  李连  张玉叶
作者单位:海军航空工程学院研究生管理大队,海军航空工程学院控制工程系,海军航空工程学院研究生管理大队 烟台,264001
摘    要:介绍了SVM方法原理,为了将SVM在解决两类分类问题中的优越性推广至解决多类分类问题,分析构建树方法与SVM结合运用来提高SVM在进行多类分类时的训练效率的方法, 然后构造文本自动分类,实验说明该方法有较高的训练效率,并且所需的训练样本量大大降低。

关 键 词:支持向量机  决策树  多类分类
修稿时间:2004年12月20

Research of SVMs Multi-Category Classification in Decision Tree
WANG Chun-xin,LI Lian,ZHANG Yu-ye Graduate Students'Brigade of NAEI.Research of SVMs Multi-Category Classification in Decision Tree[J].Journal of Naval Aeronautical Engineering Institute,2005,20(2):254-256.
Authors:WANG Chun-xin  LI Lian  ZHANG Yu-ye Graduate Students'Brigade of NAEI
Institution:WANG Chun-xin,LI Lian,ZHANG Yu-ye Graduate Students'Brigade of NAEI Department of Control Engineering,NAEI,Yantai,264001
Abstract:This paper introduces the theory of Support Vector Machine (SVM) . In order to extend the advantage of SVM in solving binary classification to multi-category classification, it analyzes the method of combining SVM with the method constructing tree to improve the training efficiency in multi-category classification. And it performs a text classification experiment to prove that it not only has high training efficiency, but also greatly reduces the number of samples.
Keywords:support vector machine (SVM)  decision tree  multi-category classification
本文献已被 CNKI 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号