基于改进证据理论的冲突证据融合方法
Conflict evidence fusion method based on improved evidence theory
张斌 1侯紫霞 1张建锋1
作者信息
- 1. 西北农林科技大学信息工程学院,陕西杨凌 712100
- 折叠
摘要
为解决证据理论中Dempster组合规则融合高度冲突证据效果不佳的问题,提出一种同时考虑证据重要性与可靠性的综合权重生成方法和基于加权复数基本概率分配函数的证据融合方法.以提高证据融合结果的收敛性能和诊断精度为目标,建立相应的优化证据融合模型.实验结果表明,所提方法能够有效识别并处理证据冲突,克服证据加权融合过程中剩余信度被完全抛弃的问题,获得更好的融合结果和更快的收敛速度,提高诊断精度.
Abstract
To address the problem of poor performance caused by highly conflicting evidence in the Dempster combination rule fu-sion of evidence theory,a comprehensive weight generation method that considered both the importance and reliability of evi-dence,and an evidence fusion method based on weighted complex basic probability assignment function were proposed.A corre-sponding optimized evidence fusion model was established with the goal of improving the convergence performance and diagnostic accuracy of evidence fusion.Experimental results show that the proposed method can effectively identify and handle evidence con-flicts,overcome the problem of completely discarding residual credibility in the evidence weighting fusion process,and achieve better fusion results and higher convergence speed,effectively improving diagnostic accuracy.
关键词
信息融合/多准则决策/证据理论/冲突证据融合/证据融合规则/证据综合权重/置信分布Key words
information fusion/multi-criteria decision making/evidence theory/conflict evidence fusion/evidence fusion rules/comprehensive evidence weight/belief distribution引用本文复制引用
基金项目
陕西省重点研发计划(2023-YBNY-212)
出版年
2024
NETL
NSTL
万方数据