基于皮尔逊系数和不确定测度的冲突证据组合方法

D-S 证据理论在合成冲突较大的证据时会产生直觉相悖问题,现有的修正证据源的改进方法大多只从单一角度做出改进,不能全面反映冲突信息特征。 针对此问题,提出了一种基于皮尔逊相关系数和不确定性测度的证据组合方法。 首先,利用皮尔逊相关系数衡量证据之间的相关性,定义证据的可信度。 其次,引入基于区间概率的不确定度对可信度进行修正,得到权重。 最后,使用该权重对原始证据进行加权平均,使用 Dempster 组合规则进行合成。 与经典改进方法相比,该方法能有效处理冲突证据的融合问题,指认正确命题的准确率达到 0. 992 0。 与已有皮尔逊系数改进方法相比该方法更具有合理性,且有较高的准确度。

Conflict evidence combination method based on Pearsoncoefficient and uncertainty measure

D-S evidence theory will produce intuitionistic conflict when synthesizing evidence with large conflict. Since most of theexisting improvement methods for correcting evidence sources only make improvements from a single perspective, they cannot fully reflectthe characteristics of conflict information. In order to solve this problem, a new evidence combination method based on Pearsoncorrelation coefficient and uncertainty is proposed. Firstly, using the Pearson correlation coefficient to measure the correlation betweenevidences, to define the credibility of the evidence. Secondly, the uncertainty based on interval probability is introduced to modify thecredibility to obtain the weight. Finally, using the weight of the weighted average of the original evidence, synthesized using Dempstercombination rule. In the case analysis, compared with the classic improved method, the proposed method can effectively deal with thefusion of conflicting evidence, and the accuracy of identifying the correct proposition reaches 0. 992 0. Compared with the existingPearson coefficient improvement method, the proposed method is more reasonable and has higher accuracy.