首页 | 本学科首页   官方微博 | 高级检索  
     检索      

一种基于不确定度的证据冲突识别方法
引用本文:李杨,郭亚军.一种基于不确定度的证据冲突识别方法[J].东北大学学报(自然科学版),2017,38(6):892-897.
作者姓名:李杨  郭亚军
作者单位:(1. 东北大学 工商管理学院, 辽宁 沈阳110169; 2. 华北理工大学 建筑工程学院, 河北 唐山063210)
基金项目:“十二五”国家科技计划项目(2013BAJ10B09-2); 河北省教育厅科研项目(QN2015183).
摘    要:从高度冲突证据对D-S证据理论的应用约束出发,针对几种常用冲突度量方法存在的问题,定义了描述证据自身特征的不确定度,将证据冲突涵义从证据间不一致扩展到证据内外,给出基于不确定度的证据冲突多元度量,并为增强冲突的识别效果提出最大差异度的冲突处理方法.验证分析得出:本文给出的度量方法能够有效表征证据间的冲突程度,最大冲突差异度的处理实现了证据冲突识别的显著化,在证据冲突度量与识别中具有很好的普适性和有效性.

关 键 词:证据冲突  多元度量  不确定度  差异度  

A Method of Evidence Conflict Identification Based on the Uncertainty Degree
LI Yang,GUO Ya-jun.A Method of Evidence Conflict Identification Based on the Uncertainty Degree[J].Journal of Northeastern University(Natural Science),2017,38(6):892-897.
Authors:LI Yang  GUO Ya-jun
Institution:1. School of Business Administration, Northeastern University, Shenyang 110169, China; 2. School of Civil and Architectural Engineering, North China University of Science and Technology, Tangshan 063210, China.
Abstract:This article defines the uncertainty degree of describing the features of evidence itself starting from the highly conflict evidence’s constraints on the application D - S evidence theory, and aiming at the existing problem of several common conflict measurement method. The connotation of evidence conflict is extended from inconsistency between the evidences to both inside and outside of evidence. Evidence conflict multi-measurement is given based on uncertainty, and the conflict processing method of maximum difference degree is put forward in order to enhance the effect of the conflict recognition. Validation analysis showed that the measurement method in this paper can effectively represent conflict degree between evidences; the processing of maximum conflict difference degree achieves significance of evidence conflict identification, which has good universality and effectiveness in evidence conflict recognition of complex systems.
Keywords:evidence conflict  multi-measurement  uncertainty degree  difference  
本文献已被 CNKI 等数据库收录!
点击此处可从《东北大学学报(自然科学版)》浏览原始摘要信息
点击此处可从《东北大学学报(自然科学版)》下载免费的PDF全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号