一种新的基于证据权重的D-S改进方法

针对D-S证据理论处理高冲突证据时会出现于直觉相悖结论的问题,提出了一种基于证据权重组合的方法。首先通过引入Jousselme距离函数来确定证据权重。然后冲突证据由平均证据代替;且其权重也赋给平均证据。最后对修正后的证据加权平均后再用Dempster组合规则实现数据融合。与其他改进方法通过实例比较,表明该方法在有冲突证据时更能有效快速地识别出目标。

A New D-S Improved Approach Based on the Weight of Evidence

Since D-S evidence theory involves counter-intuitive behaviors when dealing with highly conflict evidence. Therefore a new method based on credibility was proposed. Firstly, a distance function of Jousselme that was used to establish the credibility of the evidence was put forward. Secondly, the conflict evidence was replaced by the average evidence and their credibility was also assigned to the average evidence. Finally, after weighted averaging the modified evidence in the system, the Dempster combination rule was used to get the information fused. Compared with other improved methods by an example suggests that when the problem is of highly conflict evidence, the new approach is more efficient and quickly in identifying the target.