论文部分内容阅读
为了更好的处理运用D-S证据理论时出现高度冲突证据的情况,提出了一种基于广义Jaccard系数的加权组合方法。该方法利用广义的Jaccard系数公式求出各个证据体的权值,提高相近识别结果的证据体支持度;然后对证据体的权值归一化处理、加权平均,减少算法的复杂度;最后利用标准的D-S合成规则进行多次信息融合,进一步减小冲突证据的影响,从而给出符合实际的辨识结论。详细介绍了该方法的理论推导过程和具体计算步骤。实际案例计算表明:该方法相较于其他改进方法,在处理不相关性冲突证据方面更具有简单性、实用性以及高效性。
In order to deal with the evidence of highly conflicting evidence when using D-S evidence theory, a weighted combination method based on generalized Jaccard coefficients is proposed. The method uses the generalized Jaccard coefficient formula to calculate the weight of each evidence body and enhances the support of the evidence body of similar recognition results. Then, the weight of the evidence body is normalized, weighted average, and the complexity of the algorithm is reduced. Finally, Standard DS synthesis rules for multiple information fusion, to further reduce the impact of conflict evidence, thus giving a realistic identification conclusion. The theoretical derivation process and specific calculation steps of this method are introduced in detail. The actual case calculation shows that this method is more simple, practical and efficient in dealing with evidence of irrelevant conflict than other improved methods.