The Normalization of the Dempster’s Rule of Combination
ICAISC'10: Proceedings of the 10th international conference on Artifical intelligence and soft computing: Part II(2010)
摘要
A new method for normalization and combination of interval-valued belief structures within the framework of Dempster-Shafer theory of evidence based on the so-called ”interval extended zero” method is proposed. The two desirable intuitively obvious properties of normalization procedure are defined. The main of them is based on the assumption that the sum of normalized intervals should be an interval centered around 1 with a minimal width. The advantages of a new method are illustrated with use of numerical examples. It is shown that a new method performs better than known methods for combination of interval-valued belief structures as it provides the results with the properties which are close to the desirable ones.
更多查看译文
关键词
Interval Dempster’s rule of combinaion,Normalization
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络