基于链图的最小独立图算法研究  

Research on Minimal Independent-map of Chain Graph

作　　者：逄潇 许成 韩仃仃 PANG Xiao;XU Cheng;HAN Ding-ding(School of Mathematics and Statistics,Qingdao University,Qingdao 266071,China)

机构地区：[1]青岛大学数学与统计学院,青岛266071

出　　处：《青岛大学学报（自然科学版）》2018年第2期9-13,共5页Journal of Qingdao University(Natural Science Edition)

基　　金：山东省自然科学基金项目(批准号:ZR2016AM29)资助

摘　　要：链图的条件独立性随给定变量次序的变化而变化。定义有向图上的最小独立性并将其扩充到链图上,给定链图的一个变量次序,从而得到与给定目标次序一致的最小独立图,并给出得到该最小独立图的算法。介绍了算法的应用并讨论了其复杂度和研究前景。In order to study the conditional independence structure,introducing the concept of chain graph and studying optimization of chain graph model.Given a chain graph and a target ordering on its variables,we efficiently derive its minimal I-map whose edges are consistent with the target ordering.This paper reviews the expression of probabilistic graph model,gives the definition of conditional independence on chain graph,and then an algorithm for finding the minimum I-map is obtained,and introduces the application of the algorithm in detail.Finally,the complexity of the algorithm and the prospect of the research are discussed.

关 键 词：链图 条件独立性 最小I-map 

分 类 号：O212[理学—概率论与数理统计]