节点文献

基于Pignistic概率函数和相关系数的证据组合方法

Evidence combination method based on Pignistic function transformation and correlation coefficient

  • 推荐 CAJ下载
  • PDF下载
  • 不支持迅雷等下载工具,请取消加速工具后下载。

【作者】 杨晓萍廉伟健李孟杰钱志鸿

【Author】 YANG Xiaoping;LIAN Weijian;LI Mengjie;QIAN Zhihong;College of Communication Engineering, Jilin University;

【机构】 吉林大学通信工程学院

【摘要】 D-S证据理论不能很好地解决证据之间高冲突的情况,得到的结果与事实相悖,并且在实际的应用中无法解决"一票否决"的问题。针对上述不足,提出了一种基于Pignistic概率函数和相关系数的新证据冲突衡量标准(Pcor)和基于该标准的加权证据组合方法。该方法用Pcor来衡量证据间的冲突,通过建立支持矩阵来确定各证据的支持度,用加权平均法修正证据,使用Dempster组合规则,最终得出与事实相符的结果。数值算例表明,所提方法有效地解决了高冲突证据组合问题,与相关算法相比,具有更好的收敛性。

【Abstract】 The situation of high conflict between evidences and one ballot veto cannot be solved by classical D-S rule,and the results obtained from classical D-S rule are contrary to the facts. To solve this problem, a new standard to measure conflicts between evidences was proposed based on the combination of Pignistic function transformation and correlation coefficient, and also a novel kind of weighted combination method which was applied to measure conflicts between evidences was put forward according to the standard. After that, a support matrix was constructed based from which the credibility of evidence was obtained, and the weighted average method was used to revise the evidence. Finally, the combination was accomplished by using Dempster’s rule. The result of numerical examples shows that it’s effective to solve the combination of conflicting evidence. Compared with other methods, the proposed method has good astringency.

【基金】 吉林省校共建计划专项基金资助项目(No.SXGJQY2017-9);吉林大学高层次科技创新团队建设基金资助项目(No.2017TD-19);国家自然科学基金资助项目(No.61771219);吉林省科技厅重点科技研究发展基金资助项目(No.20180201081SF)~~
  • 【文献出处】 通信学报 ,Journal on Communications , 编辑部邮箱 ,2019年03期
  • 【分类号】TP202;TP212
  • 【被引频次】20
  • 【下载频次】276
节点文献中: 

本文链接的文献网络图示:

本文的引文网络