一种基于证据间距离和信息熵的证据组合方法  

Combination Method of Eidence Based on Distance of Evidence and Information Entropy

作　　者：金晓斌[1] 许大琴[1] 谈亮[1] JIN Xiao-bin;XU Da-qin;TAN Liang

机构地区：[1]海军指挥学院,江苏南京211800

出　　处：《指挥控制与仿真》2016年第6期113-117,共5页Command Control & Simulation

基　　金：军事类研究生资助课题(2014JY437)

摘　　要：针对改进的基于证据间距离的组合方法,首先计算证据间距离并求出各个证据的支持度和可信度,选取支持度高的作为较可信证据,其次引入证据信息熵,计算修正系数,对这些证据的可信度进行修正,将可信度作为权重,进行证据的加权平均,再利用Dempster组合规则对所有证据进行融合。算例表明,该方法比现有的方法更加有效。In order to combine highly conflict evidence more efficiently,an improved method based on the distance of the bodies of evidence was proposed. Firstly,it computed the distance between the bodies of evidence and obtained the support degree and the credibility of each evidence.Then,information entropy was introduced,the correction factor was computed and the credibility of the evidence which had better support degree was corrected. Finally,the credibility was taken as weights and weighted averaging all the evidence,and the Dempster combination rule was used to realize information fusion.The validity of the approach was demonstrated by simulation example.

关 键 词：证据理论 冲突证据 证据距离 信息熵 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] E911[自动化与计算机技术—计算机科学与技术]