基于皮尔逊系数的冲突证据合成新方法

提出了一种融合证据源预处理和合成规则修改的冲突证据合成方法.算法得到两两证据源间的皮尔逊系数,每个证据源具有特有皮尔逊系数值和.以具有最大皮尔逊值和的证据源为中心证据源,所有证据源相比中心证据源的比例为该证据源的权值,权值反映了证据的重要程度,从而冲突证据源具有最小的权值.利用权值对原始证据源进行预处理,降低冲突证据的重要性.再通过改进的合成规则,对处理后的证据源进行合成,得到最终合成结果.证据正常与冲突情况下的仿真算例验证了新方法不但在证据合成中具有很高的通用性,而且相比其他方法具有很强的优越性.

A novel D-S combination method of conflicting evidences based on Pearson correlation coefficient

Based on the improvement of both evidence sources and combination rules, a novel conflict evidence combination method is proposed. Pearson coefficients between evidences are calculated. The evidence with maximum sum value of Person coefficients is chosen to be central evidence. The weights of evidence sources are obtained based on the proportion compared to the central evidence, which reflect the importance of the evidence. According to the weights, all evidence sources are pretreated. After that, the final combination results are obtained by modified combination rules. Examples with evidences of normal and conflict situations verify that the new method has not only high versatility in the evidence combination, but also great superiority compared with the other methods.