概率模糊认知图  

Probabilistic Fuzzy Cognitive Maps

作　　者：骆祥峰[1] 高隽[1] 

机构地区：[1]合肥工业大学计算机与信息学院图像信息处理研究室,安徽合肥230009

出　　处：《中国科学技术大学学报》2003年第1期26-33,共8页JUSTC

基　　金：国家自然科学基金项目 (No .60 175 0 11);安徽省自然科学基金 (No .0 10 42 3 0 1);教育部优秀青年教师资助计划项目资助

摘　　要：模糊认知图模型通过在传统认知图模型中引入模糊测度来量化概念间因果关系的影响程度 .在继承模糊认知图模型优点的前提下 ,首次在概念间的因果关系中引入条件概率测度 ,提出概率模糊认知图模型 .该模型不仅能表示概念间的定性及模糊因果关系 ,而且能表示概念间的条件概率因果关系 ,并能退化为模糊认知图 .通过对复杂机械部件的拆卸与装配实验 ,具体阐述概率模糊认知图的应用 .Kelly considered that there exist only increase an d decrease relationships between states of concepts in cognitive maps. The relatio nships between the cause and the effect concepts in kelly's cognitive map can be extended into fuzzy relationships by fuzzy cognitive map (FCM). In the viewpoin t of the authors, there does exist not only fuzzy relationships but also probabi listic relationships in the real world. A new cognitive map-probabilistic fuzzy cognitive map (PFCM) is presented based on the idea. In PFCM, relationships bet ween concepts can be binary set, fuzzy set or conditional probability, and the a dvantages of cognitive maps are preserved. To illustrate the application of PFCM , disassembling and assembling process of mechanical parts are simulated. The ex periment results show that PFCM is more flexible and robust than FCM.

关 键 词：模糊认知图 概率模糊认知图 模糊因果关系 概率因果关系 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]