一种最优证据选择方法  

An Optimal Evidence Selection Approach

作　　者：李文艺[1] 吕现钊[1] 岳明道[1] 

机构地区：[1]宿州学院机械与电子信息工程学院,安徽宿州234000

出　　处：《蚌埠学院学报》2014年第1期24-26,共3页Journal of Bengbu University

基　　金：安徽省教育厅校企合作项目(KJ2013Z316);宿州学院硕士科研启动基金项目(2009yss08;2009yss07)

摘　　要：给出了一种从多个证据中选择最优证据的方法,该方法能解决证据冲突问题。给出了证据量、证据离散度、证据冲突半径等概念。首先构造出冲突矩阵,再计算出每个证据的冲突半径,按照冲突半径从小到大对证据进行排序。最后,对排序后的证据按顺序进行分步合成,保存每一步的合成结果,并计算每一步结果的证据离散度。当离散度达到最大时,参与合成的一组证据为最优证据。实验表明该方法可以解决证据冲突问题。An optimal evidence selection approach is presented, and this method could solve the problem of evidence conflict. The definition of evidence intensity, evidence discrete degree and evidence conflict radius are provided. Firstly a conflict matrix is constructed, and all evidence conflict radiuses are calculat-ed. The evidences are sequenced according to conflict radius from small to big. At last, the sequenced evi-dences are combined one by one and the combination results are saved and each evidence discrete degree is calculated. When the evidence discrete degree comes to the maximum, the combined evidences are the optimal evidences. Experiment show that this method could solve the evidence conflict.

关 键 词：D-S证据理论 证据冲突 最优证据 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]