计算机工程与应用 ›› 2019, Vol. 55 ›› Issue (4): 154-158.DOI: 10.3778/j.issn.1002-8331.1711-0172

• 模式识别与人工智能 • 上一篇    下一篇

一种有效折扣证据源的冲突证据合成方法

罗  兰,肖建于   

  1. 淮北师范大学 计算机科学与技术学院,安徽 淮北 235000
  • 出版日期:2019-02-15 发布日期:2019-02-19

Conflict Evidence Synthesis Method of Effective Discount Evidence Sources

LUO Lan, XIAO Jianyu   

  1. School of Computer Science and Technology, Huaibei Normal University, Huaibei, Anhui 235000, China
  • Online:2019-02-15 Published:2019-02-19

摘要: 为解决证据组合规则中一票否决和弱决策证据在低冲突情况下出现的反直观推理的现象,提出一种有效折扣证据源的冲突证据合成方法。首先根据证据的信任函数和似然函数求得证据间的相似度,然后求出证据之间的支持程度,并确定折扣因子,最后采用Dempster组合规则合成利用折扣因子修正后的证据源。数值算例分析结果表明,改进后的证据组合方法可以有效地处理证据冲突。

关键词: D-S证据理论, 证据冲突, Vague集, 折扣因子

Abstract: In order to solve the phenomenon of one vote veto and weak decision evidence in anti conflict reasoning under low conflict situations, a conflict evidence synthesis method of effective discount evidence sources is proposed. Firstly, according to the belief function and likelihood function of evidence, the similarity between evidences is obtained. Then, the degree of support between evidences is obtained, and the discount factor is determined. Finally, the Dempster combination rule is used to synthesize the evidence source corrected by discount factor. The results of numerical examples validate that the improved evidence combination method can effectively deal with the conflict of evidence.

Key words: D-S evidence theory, evidence conflict, Vague set, discount factor