基于证据间距离和不确定性度量的证据组合方法

针对Dempster组合公式无法组合冲突证据的问题,提出了一种证据组合的权重分配方法.该方法充分考虑证据间的关系和证据本身的特性,用证据间距离度量证据间的不一致程度,用证据的不确定度来度量证据本身的不确定性;在此基础上扩展了文献7]提出的权重确定准则,认为证据组合规则既要考虑使组合后证据与各源证据间的距离和尽量小,也要注重降低组合后证据本身的不确定性.最后根据新的准则给出了权重因子的确定算法和证据组合方法.算例表明,该方法改进了文献7]权重分配方法的结果,且使权重分配更加灵活.

Evidence Combination Approach Based on Distance between Evidences and Ambiguity Measure

In order to solve the problem that conflict evidences can not be combined with Dempster''s combination rule, a weights redistribution approach to evidence combination is proposed. In this approach the disaccord of evidences and the ambiguity of evidence are measured by distance and ambiguity measure respectively. Based on this, this paper expands the weights determination principles proposed in \7\]. Then, new weights determination algorithm and evidence combination approach are suggested, and the result in \7\] is improved.