计算机工程与应用 ›› 2007, Vol. 43 ›› Issue (8): 29-31.

• 学术探讨 • 上一篇    下一篇

基于因果语义定向的贝叶斯网络结构学习

王双成 张明 陈乃激   

  1. 上海立信会计学院信息科学系 上海 201620
    上海立信会计学院中国立信风险管理研究院 上海 201620

  • 收稿日期:2006-07-12 修回日期:1900-01-01 出版日期:2007-03-11 发布日期:2007-03-11
  • 通讯作者: 王双成

Learning Bayesian Networks Structure Based on Causal Semanitics Orienting

ShuangCheng Wang   

  1. Department of Information Science, Shanghai Lixin University of Commerce, Shanghai 201620
    Risk Management Research Institute, Shanghai Lixin University of Commerce, Shanghai 201620
  • Received:2006-07-12 Revised:1900-01-01 Online:2007-03-11 Published:2007-03-11
  • Contact: ShuangCheng Wang

摘要: 基于变量之间基本依赖关系、基本结构、d-separation标准、依赖分析思想和混合定向策略,给出了一种有效实用的贝叶斯网络结构学习方法,不需要结点有序,并能避免打分-搜索方法存在的指数复杂性,以及现有依赖分析方法的大量高维条件概率计算等问题。

关键词: 贝叶斯网络, 结构学习, 依赖分析, 因果语义, 碰撞识别

Abstract: A new method of learning Bayesian network structure based on basic dependency relationship between variables, basic structure between nodes, d-separation criterion, the idea of dependency analysis and the strategy of mixture orienting. This method do not need sorting nodes. It can effectively avoid the exponential complexity of search & scoring based methods and a large number of the calculate of high rank conditional probability in existing dependency analysis based methods.

Key words: Bayesian networks, structure learning, dependency analysis, causal semanitics, collider identification