Computer Engineering and Applications ›› 2008, Vol. 44 ›› Issue (29): 21-23.DOI: 10.3778/j.issn.1002-8331.2008.29.005

• 博士论坛 • Previous Articles     Next Articles

Node order of Bayesian network based on feature selection using Support Vector Machine

LV Shi-pin1,2,WANG Xiu-kun1,SUN Yan2,3,TANG Yi-yuan2   

  1. 1.Department of Computer Science and Engineering,Dalian University of Technology,Dalian,Liaoning 116023,China
    2.Institute of Neuroinformatics,Dalian University of Technology,Dalian,Liaoning 116023,China
    3.Department of Computer & Information Technology,Liaoning Normal University,Dalian,Liaoning 116029,China
  • Received:2008-06-25 Revised:2008-07-21 Online:2008-10-11 Published:2008-10-11
  • Contact: LV Shi-pin

基于支持向量机特征选择的贝叶斯网结点序

吕世聘1,2,王秀坤1,孙 岩2,3,唐一源2   

  1. 1.大连理工大学 计算机科学与工程系,辽宁 大连 116023
    2.大连理工大学 神经信息学研究所,辽宁 大连 116023
    3.辽宁师范大学 计算机与信息技术学院,辽宁 大连 116029
  • 通讯作者: 吕世聘

Abstract: At present,the method of search and score is widely used for learning the structure of Bayesian network.The method needs first the node order in the network,which is usually decided according to user’s experience,so the strong subjectivity blocks the method’s practical application.By measuring every node’s influence on the leaf node,feature selection based on support vector machine can learn the node order from data and get rid of the effects of human factors.Experimental results show the proposed method is effective.

Key words: Bayesian network, support vector machine, feature selection

摘要: 目前较常采用搜索打分方法进行贝叶斯网络结构学习,该方法需要首先依据参与者的经验来确定网络的结点顺序,主观性较强,限制了它的实际应用。基于支持向量机特征选择的方法,可以按照各个结点对叶结点的影响能力进行排序,从而直接从数据中通过学习得出结点顺序,避免了人为因素的影响。实验结果验证了该方法的有效性。

关键词: 贝叶斯网络, 支持向量机, 特征选择