摘要
由于多证据源各有擅长,即使同一证据源对于不同事件判定的置信度也各不同,使用传统Dempster-Shafer理论往往得到与直觉相悖的结果,因此使用不同的擅长因子衡量证据对不同焦点元素判定的置信度,并将其引进信任函数,所有的证据经过擅长评估后再进行组合,实验证明,给定方法比传统D-S理论提高了融合的准确性。
The wrong results are often got using traditional Dempster-Shafer rules due to the difference of credibility from multi-evidence. The evidence sources are adapted in different cases. The credibility of the evidence changes by the different focus element. So, in evidence, various credibility is given for evcry focus element. These parametersjoin in the belieffunction. Evidences are first estimated with these parameters and then combined. The conclusion is significant after the final experimental test comparing with traditional D-S theory.
出处
《计算机工程与设计》
CSCD
北大核心
2007年第12期2928-2929,共2页
Computer Engineering and Design
基金
国家科技攻关基金项目(2004-BA608B-030303)
广州番禺科技攻关基金项目(2004-IF31)
关键词
信息融合
证据理论
组合规则
入侵检测
融合监控
information fusion
evidence theory
combination rule
intrude detection
fusion monitoring