Approach to evidence combination based on fuzzy theory and its applications
Guo Hui-xin · Kongzhi yu juece · 2008
A combination rule is proposed in order to efficiently combine highly conficting evidences.The conflict weights of every two evidences in evidence set are calculated,and a conflict matrix is composed of the calculated conflict weights.A method for distinguishing the evidences bringing on conflicts is presented according to the conflict matrix.Based on fuzzy theory,the fuzzy credibility degree of evidence is defined and formulated as a function of the conflict weights.Each of evidences is preprocessed according to its fuzzy credibility degree,and then the Dempster's rule of combination is used to combine the preprocessed evidences.Examples show that the proposed approach can efficiently eliminate the influence of conflicts between evidences on evidence combination and can achieve a reliable combination result with rationality.