计算机工程与应用 ›› 2011, Vol. 47 ›› Issue (28): 56-58.

• 研究、探讨 • 上一篇    下一篇

BNs中基于分层排序的条件独立分析与应用

辛国福,杨有龙,施轶青   

  1. 西安电子科技大学 理学院,西安 710071
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2011-10-01 发布日期:2011-10-01

Analysis of conditional independence and applications based on Layer sorting in BNs

XIN Guofu,YANG Youlong,SHI Yiqing   

  1. School of Science,Xidian University,Xi’an 710071,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2011-10-01 Published:2011-10-01

摘要: 贝叶斯网络(BN)是图论和概率论有机融合的概率图形模型。D-分割(directional separation)和ud-分割(unidirectional separation)是判断贝叶斯网络中条件独立的两个重要的准则。讨论了贝叶斯网络中d-分割和ud-分割的关系,得到d-分割是ud-分割的充分不必要条件。介绍分层排序的定义并给出了d-分割和ud-分割同时成立的条件。通过对贝叶斯网络的节点进行分层排序,能够得到该贝叶斯网络的拓扑序列并且能够快速找到d-分割和ud-分割集识别有向因果效应。

关键词: d-分割, ud-分割, 贝叶斯网络, 分层排序

Abstract: BNs (Bayesian Networks) are probabilistic graphical models combining probability theory and graph theory.Directional separation and unidirectional separation are two important rules on the BNs to confirm the conditional independence.The relationship between directional separation and unidirectional separation in BNs is discussed in detail.The conclusion that directional separation is the sufficient condition for unidirectional separation is obtained.The definition of layer sorting is introduced and the condition that both directional separation and unidirectional separation hold is proposed.By using layer sorting the nodes of a BN,it can get a BN’s topological sequence and find d-separation and ud-separation sets to indentify direct causal effect quickly.

Key words: directional separation, unidirectional separation, Bayesian Networks(BN), layer sorting