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,已经发展成了一个能够处理离散变量和连续变量的混合模型。但已有的因果图的推理算法还不能完全适应实际问题的需要,这大大地限制了因果图推广和使用,然而信度网研究已比较成熟,已有许多现成的算法和实用的推理软件。文中给出了从因果图向信度网转化的一般方法,包括因果图的连接强度向信度网的条件概率表转化和因果图的结构向信度网的结构转化,从而可以利用信度网的这些成果。
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 the reasoning algorithms of Causality Diagram are not so many that they limit the popularization and application of Causality Diagram. The Belief Network has many ready reasoning algorithms and applied software. This paper presents a method of how to transfer from Causality Diagram to Belief Network. It includes the transfer of the linkage intensity of Causality Diagram to CPT of Belief Network, and the transfer of the structure of Causality Diagram to the structure of Belief Network.
出处
《计算机仿真》
CSCD
2004年第10期89-92,共4页
Computer Simulation
基金
重庆市科技攻关项目(5990)
(7359)
重庆市科委项目基金资助项目(91214297)资助。
