基于相关系数与相关距离的证据合成方法

摘要

针对传统Dempster-Shafer证据合成方法合成冲突证据时,会出现相悖的结论的问题,提出了基于相关系数与相关距离的冲突证据合成方法.方法利用证据冲突信息,通过相关系数与相关距离得到证据距离,距离表征证据的冲突程度,并进一步计算得到权重系数,根据权重系数对证据进行重新概率分配后,完成证据预处理过程,最后用改进的合成规则,完成再分配证据的合成.具体的证据实证验证了算法在合成冲突证据方法具有最优结果,且和推理一致,可以很好的用于证据合成.%For the conflict evidence combination,the Dempster-Shafer evidence combination method will appear the problem of contrary conclusion,and so the method of conflict evidence synthesis based on correlation coefficient and correlation distance was proposed.According to the angle cosine concept,the multi-dimensional vector distance formula was inferred.After construction of the evidence vector,evidence distances which represent the conflict degree were calculated,and which were further converted into the weight coefficient.The probability distributions of evidences were redistributed by using the weight coefficient,so the evidence pretreatment process was completed.After that,using the improved synthesis rules,the synthesis of the evidence of redistribution was completed.The empirical results show that the algorithm has the best results in the conflict evidence synthesis,and is consistent with the reasoning,which can be used for the evidence synthesis.