基于遗传算法的因果图网络结构学习  

Learning Causality Diagram Structure Based on Genetic Algorithm

作　　者：石庆喜[1] 梁新元[1] 张勤[2] 

机构地区：[1]重庆工商大学计算机学院,重庆400067 [2]重庆大学自动化学院,重庆400030

出　　处：《重庆大学学报（自然科学版）》2006年第4期111-114,共4页Journal of Chongqing University

基　　金：重庆市科技攻关资助项目(5990)

摘　　要：在因果图理论中,采用了图形化和直接因果强度来表达知识和因果关系,它克服了贝叶斯网的一些不足,已发展成了一个能够处理离散变量和连续变量的混合模型.但是因果图的结构得由领域专家给出,这在实际中很难办到.鉴于因果图结构的复杂度随论域中节点个数的增加呈指数上升,寻找最有可能因果图网络结构成为了NP-HARD难题.文中给出了如何利用已知数据集,寻找最有可能的因果图网络结构设计的遗传算法(Genetic A lgorithm,GA).The Causality Diagram theory, which adopted graphical expression of knowledge and direct causality intensity of causality, overcomes some shortages in Belief Network and has evolved into a mixed causality diagram methodology coped with discrete and continuous variable. But it is difficult that the structure of Causality Diagram given by expert. Because the complexity of causality diagram structure goes up exponentially through the number of the vertex＇ s increasing, it is NP-hard problem to find the most possible structure from a set of data. The authors discuss approaches and present Genetic Algorithm, to find the most possible structure from a set of data. Experiment shows the method is effective.

关 键 词：因果图 因果图网络结构 机器学习 遗传算法 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]