摘要
为了解决现有对于高度冲突证据的合成方法存在的问题与不足,提出了一种新的合成方法.新方法以相似性质的冲突程度和相似系数为基元,利用改进的KL来计算得到证据距离,通过证据距离和中心证据的关系建立权值,权值反映了证据的重要程度,在已有的基本可信度分配函数中加入权值进行修正.最后通过改进的方法,得到合成结果.通过算例比较,表明新方法的合成结果更为理想,更加符合实际情形.
Against the existing problems and shortage of conflict evidence combination method, one new improved method was studied. The conflict degree and similar coefficient have similar property, which both show the correlation of evidence, the evidence distance was calculated using them based on KL equation. According to the distance between with the central evidence, each evidence weight which reflects the importance of evidence was calculated, and then the existing distribution functions was amended. The final result was gotten by the improved combination methods. Example of comparison with other algorithm shows that the new method is more consistent with the actual situation.
出处
《四川大学学报（自然科学版）》
CAS
CSCD
北大核心
2013年第3期499-502,共4页
Journal of Sichuan University(Natural Science Edition)
基金
中国民用航空总局科技项目(MHRD200925)
关键词
数据融合
证据合成
冲突
证据距离
权值
data fusion
evidence combination
conflict
evidence distance
weight
