摘要
Dempster-Shafer证据理论目前已被广泛应用于大数据时代的各行各业,但是当该理论应用在高度冲突的证据源进行融合时,往往会产生一些有悖直觉的结果,具有一定的局限性.为了解决这一问题,该文提出了一种基于双重中位数绝对偏差(MAD)检测和一种新的证据加权组合改进方法.首先通过MAD算法检测出异常证据,再使用这组证据的平均值对异常证据进行修正,然后使用新的证据加权组合方法对证据加权平均后得到最终结果,最后通过随机模拟实验和具体的算例实验,并与其他几种经典的证据合成规则进行比较.仿真结果表明该文提出的方法能有效地解决高度冲突的证据融合问题.
Dempster-shafer evidence theory has been applicable in all walks of life of digital marketing in a large scale,but when it is applied to the integration of highly conflicting evidence sources,it often produces some counterintuitive results and has certain limitations.In order to solve this problem,the improved method based on double median absolute deviation detection and a new improved method of evidence weighted combination evidence weighted combination is proposed.Firstly,the abnormal evidence is detected by MAD algorithm,then the average value of this group of evidence is used to correct the abnormal evidence,and the new evidence weighting combination is applied to weight average the evidence to get the final count.Through random simulation experiments and example experiments,and compared with several other classical evidence synthesis rules,the simulation results show that the synthesis rules in this paper can effectively solve the problem of highly conflicting evidence fusion.
作者
余鑫
吴根秀
蔡奥丽
YU Xin;WU Genxiu;CAI Aoli(College of Mathematics and Statistics,Jiangxi Normal University,Nanchang Jiangxi 330022,China)
出处
《江西师范大学学报（自然科学版）》
CAS
北大核心
2022年第5期523-532,共10页
Journal of Jiangxi Normal University(Natural Science Edition)
基金
国家自然科学基金(61876074)资助项目.
关键词
D-S证据理论
中位数绝对偏差
高度冲突
Dempster-Shafer evidence theory
median absolute deviation
high conflict
