基于证据支持一致性分类的证据合成方法  

Evidence Combination Method Based on Evidence Support Consistency Classification

作　　者：帅飞[1] 李俊全[1] 万伟[2] 

机构地区：[1]解放军信息工程大学三院,郑州450004 [2]北京航天飞行控制中心,北京100094

出　　处：《计算机工程》2012年第21期123-125,共3页Computer Engineering

摘　　要：利用D-S证据理论评估信息安全产品的安全等级时可能产生全冲突悖论、1冲突悖论、0冲突悖论。为解决该问题,提出一种基于证据支持一致性分类的证据合成方法。基于证据支持一致性对证据进行分类后,根据专家意见的重要程度,对证据进行折扣,并将D-S合成方法和加权平均方法相结合,用于合成分类证据。分析结果证明,该方法有更好的适用性和更低的计算复杂度。When D-S evidence theory is applied to assess the security level of information security products, it may produce the whole conflict paradox,1 confict paradox and 0 confict paradox. To solve this problem, this paper designs an evidence combination method based on evidence support consistency classification. It classifies the evidence according to support consistency, discounts evidence according to the degree of importance of expert advice, and combines D-S combination method and weighted average method to combine classified evidences. Analysis results show that the method has higher applicability and lower computational complexity.

关 键 词：安全等级评估 D—S证据理论 支持一致性 证据冲突 证据合成 专家决策 

分 类 号：TP309.1[自动化与计算机技术—计算机系统结构]