计算机工程与应用 ›› 2010, Vol. 46 ›› Issue (10): 36-37.DOI: 10.3778/j.issn.1002-8331.2010.10.012

• 研究、探讨 • 上一篇    下一篇

改进的不均衡样本集支持向量机预处理方法

何渊淘,邓 伟   

  1. 苏州大学 计算机科学与技术学院,江苏 苏州 215006
  • 收稿日期:2008-10-08 修回日期:2009-01-15 出版日期:2010-04-01 发布日期:2010-04-01
  • 通讯作者: 何渊淘

Improvement on preprocessing algorithm of support vector machines for unbalance data set

HE Yuan-tao,DENG Wei   

  1. Department of Computer Science and Technology,Soochow University,Suzhou,Jiangsu 215006,China
  • Received:2008-10-08 Revised:2009-01-15 Online:2010-04-01 Published:2010-04-01
  • Contact: HE Yuan-tao

摘要: 将一种改进的邻域算法应用于不均衡样本集中,由于改进的邻域算法未考虑不均衡样本集的问题从而导致后续的支持向量机训练耗费和泛化性能受影响,把后验概率的思想加入改进的邻域算法中,并由实验数据说明了该方法对不均衡样本集的有效性。

关键词: 支持向量机(SVM), 邻域算法, 后验概率, VC维

Abstract: An improved neighborhood algorithm is applied in unbalance data set.Because of not considering unbalance date set,the generalization of improved neighborhood algorithm is degrade.In order to solve this problem,the idea of posterior probability is used in the new neighborhood algorithm and the result of pattern classification shows the effectiveness of the method.

Key words: Support Vector Machine(SVM), neighborhood algorithm, posterior probability, VC dimension

中图分类号: