凸函数证据理论模型

A CONVEX EVIDENCE THEORY MODEL

吉林大学知识工程研究小组针对有序命题类问题给出了适合于有序命题类问题的一类证据理论模型，该模型中，由于ｇ 的计算不够准确导致了组合误差，最大误差为一个分类等级．文中通过给出计算ｇ 的新方法予以解决．此外在证据理论中，通常不同的Ｍａｓｓ 函数代表不同专家对某一事物的看法，一般说来不同专家的权威性不尽相同，但迄今为止证据理论方面的研究还未给出组合两个具有不同权威性的Ｍａｓｓ 函数的方法。

To solve the problems of the ordered propositions, an evidence theory model is given. But this model produces a combinational error because of computing precision. This problem is solved by presenting a new method of calculating parameter g . On the other hand, the different Mass functions denote the varied experts' options for the same thing, and the authorities of varied experts are different. But until now, There does not exist a method which combines two different weighted mass functions. Based on the above research results, this paper provides a method which can combine two different weighted mass functions.