摘要
因果图理论是利用图形化和直接因果强度来表达知识和因果关系的一种基于概率论的推理方法,能够进行在线动态推理和对复杂系统进行故障诊断,连接强度是其推理的基础,文中给出了采用EM(η)算法在线学习因果图参数(连接强度)的方法,使学习出的参数能适应环境的变化,具有适时性.同时在理论上证明了这种方法的可行性和优点.
Causality diagram theory is a methodology based on probability, which adopts graphical expression of knowledge and direct causal intensity of causality, it can progress dynamic reasoning online and fault diagnosis for complex system. Linkage intensity is the basis of the inference. The algorithm of EM(η) is proposed to learn causality diagram parameters (linkage intensity), which can make the parameters adapt with the change of environment. The authors also prove the feasibility and advantage in theory.
出处
《重庆大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2006年第3期84-86,共3页
Journal of Chongqing University
基金
重庆市高等学校优秀中青年骨干教师资助项目(渝教人2005.02)
重庆市科技攻关资助项目(5990)
关键词
因果图
信度网
连接强度
EM(η)算法
causality diagram
belief network
linkage intensity
EM(η) algorithm