摘要
为了有效融合高度冲突的证据,在三角模算子和折扣因子分析的基础上,提出了一种基于证据分类的冲突证据融合规则。采用基于3角模算子定义的平均证据距离与冲突因子将证据分成可信任证据、不冲突证据和冲突证据三类,并赋予可信任证据和不冲突证据折扣因子1,极大程度上保留了证据对正确假设的支持;然后基于证据距离定义了改进的证据权重,基于加权原则对冲突证据进行合成得到修正的证据体,从而消除证据间的冲突;最后利用Dempster规则完成证据组合。算法分析表明所提方法是合理有效的。
In order to combine highly conflicting evidence efficiently,a new evidence combination rule making use of evidence classfication was proposed based on triangular norm and discount factor analyse.First,utilizing average evidence distance and discount factor based on triangular norm,the evidence was classified into three categories：reliability,no conflict and conflict.The discounting factors of the former two categories of evidences were set to one,which keeps the evidence hold of the right hypothesis to a great extant and makes the fusion results focus onto the right hypothesis more strongly.Then the was improved evidence weight was obtained based on evidence distance,and the modified evidence was obtained by verifying the conflicting evidence according to the weighting rule in order to eliminate the conflict.Finally,according to the Dempster＇s rule,the modified evidence was combined.Numerical examples show the efficiency and rationality of the proposed approach.
出处
《计算机科学》
CSCD
北大核心
2013年第1期247-250,共4页
Computer Science
基金
国家自然科学基金项目(61263031)
甘肃省自然基金项目(1010RJZA046)
甘肃省教育厅研究生导师基金项目(0914ZTB003)资助
关键词
证据理论
冲突
折扣因子
三角模算子
Evidence theory
Conflict
Discount factor
Triangular norm
