Combination Of Conflicting Evidences Based On Set Covering Model

Uncertainty reasoning can use relevant knowledge to infer a reasonable or nea.r-reasonable conclusion in the case of informa.tion lacking.D-S evidence theory is an important method in the fields of uncertain reasoning,multi-source information fusion and decision analysis.It is based on bayesian theory,and proposed by Dempster and his student Shafer.However,the fusion result may be contrary to intuition when the D-S evidence theory is used to deal with conflicting evidence.This issue has been widely concerned by researchers.Many scholars have improved the fusion method proposed by Dempster and Shafer,such as Murphy method,Ya.ger method,DP rule and so on.Based on the existing resea.rch,the following work has been done in this paper:(1)A method called Combination of Conflicting Evidences Based on Covering Model(CCEM)is proposed.This method starting from existing research applies set covering theory to cover the original information system with several non-conflicting sub-information systems,and then synthesizes each sub-information system through evidence theory.Finally,the weight of each subsystem is defined by using the idea of concept support and is applied to synthesis the probability assignment function.(2)The validity of CCEM is verified.(3)Based on the research of the CCEM,the attribute reduction that keeps the expert consistent space unchanged is defined and the relevant heuristic algorithm is given.CCEM is different from existing methods in conflict evidence processing.Firstly.CCEM covers the original D-S evidence theory.That is to say,D-S evidence theory is a special case of CCEM.Furthermore,CCEM can not only deal with the information system with conflicts,but also deal with the information system without conflicts.And then,each object needs to be compared when constructing the non-conflicting subsystem.So CCEM pays more attention to the connection between objects in the information system and makes full use of the information table.