DocumentCode :
2256670
Title :
Combination approach for conflicting evidence based on discount coefficient
Author :
Zhao, Yuxin ; Jia, Renfeng
Author_Institution :
College of Automation, Harbin Engineering University, Harbin 150001, P.R. China
fYear :
2015
fDate :
28-30 July 2015
Firstpage :
4841
Lastpage :
4845
Abstract :
In order to overcome the drawback of Dempster´s combination rule in fusing the conflicting evidence, a new combination approach based on discount coefficient is proposed. The support degree among the pieces of evidence, the certainty degree and the decision degree of the evidence itself were used to calculate the weight of each piece of evidence. The discount coefficient was obtained through the weight of the conflicting evidence and the average weight. The conflicting evidence can be modified with the discount coefficient. The fusion result was acquired by fusing the modified evidence and other pieces of credible evidence with Dempster´s combination rule. The new approach was compared with other methods by numerical examples. The results show that the proposed approach can deal with the conflicting evidence effectively, and it is more helpful to make decision.
Keywords :
Accuracy; Correlation; Entropy; Random variables; Reliability theory; Sun; combination rule; conflicting evidence; credible degree; discount coefficient;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Control Conference (CCC), 2015 34th Chinese
Conference_Location :
Hangzhou, China
Type :
conf
DOI :
10.1109/ChiCC.2015.7260389
Filename :
7260389
Link To Document :
بازگشت