作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2009, Vol. 35 ›› Issue (19): 178-180. doi: 10.3969/j.issn.1000-3428.2009.19.059

• 人工智能及识别技术 • 上一篇    下一篇

一种改进的主动支持向量机算法及其应用

樊继伟,李朝锋,吴小俊   

  1. (江南大学信息工程学院,无锡 214122)
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2009-10-05 发布日期:2009-10-05

Improved Active Support Vector Machine Algorithm and Its Application

FAN Ji-wei, LI Chao-feng, WU Xiao-jun   

  1. (School of Information Engineering, Jiangnan University, Wuxi 214122)
  • Received:1900-01-01 Revised:1900-01-01 Online:2009-10-05 Published:2009-10-05

摘要: 针对支持向量机中分类器易受样本孤立点影响的问题,提出一种改进的主动支持向量机算法,采用K-means算法获取少量“代表性”样本作为训练样本,通过训练该标识样本得到一个初始分类器,利用主动学习策略选择最佳未标记样本进行类别标记,并加入训练样本集重新训练分类器,重复该过程直到满足某些要求。运用Iris数据和遥感数据对其进行测试,实验结果表明,该算法是有效的。

关键词: K-means算法, 支持向量机, 主动学习

Abstract: Aiming at the problems that the classifier in Support Vector Machine(SVM) is influenced by sample isolated points easily, an improved active SVM algorithm is proposed, which uses K-means algorithm to obtain less representative sample as training sample. By training it, an initial classifier is got. The type of best unlabled sample is identified by means of active learning strategy, which is added into training sample sets to train classifier again. This paper repeats the process until it meets some requirements. It is tested with Iris data and remote sensing data. Experimental results show this algorithm is effective.

Key words: K-means algorithm, Support Vector Machine(SVM), active learning

中图分类号: