不完备识别框架下的证据组合方法  

The Combination of Evidences in an Incomplete Frame of Discernment

作　　者：曾成[1] 赵保军[1] 何佩琨[1] 

机构地区：[1]北京理工大学信息科学技术学院电子工程系,北京100081

出　　处：《电子与信息学报》2005年第7期1043-1046,共4页Journal of Electronics & Information Technology

基　　金：国家863计划(2002AA813032)资助课题

摘　　要：证据理论应用中经常面临识别框架不完备问题。本文引入证据框架概念,把识别框架分为已知和未知两个部分,提出一种在识别框架不完备条件下表示和组合证据的开放识别框架方法。该方法能明确区分“识别框架不完备”与“证据冲突”两个不同概念,并合理解决了对未知命题集合的基本概率赋值问题,从而克服了D-S理论以及Smets的“开放世界假设”方法在处理不完备识别框架时存在的不足。The incompleteness of the frame of discernment is one of the main difficulties confronted in the application of the evidence theory. In this paper, through introducing a newly defined evidence frame which divides the frame of discernment into known and unknown parts, an open frame method to represent and combine evidence under incomplete frame is proposed. This method can distinguish the 'incompleteness of frame' from the 'conflict of evidence' clearly. It also presents a rational way to assign a BPA to the set of unknown propositions. Therefore such a method avoids the problems confronted by the D-S theory and Smets's 'open world assumption' in dealing with incomplete frame.

关 键 词：D-S理论 证据组合 开放识别框架 证据框架 开放世界假设 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]