摘要
基本可信度数值之间的关系有时不能反映证据之间实际的联系,仅仅依靠这种数值关系来修正证据可能会出现不符实际的结论,同时也会降低合成结果的可解释性。针对该问题,面向具体决策问题提出了一种基于证据源特征的证据分组合成的方法:在经典证据框架的基础上增加了反映证据源特征的第二维信息框架,并基于第二维框架确定证据分组标准和证据修正系数,利用不同的合成规则对证据组内部和外部的信度进行修正和合成。最后,以证券市场上多位证券分析师预测信息融合的实际算例,验证了该方法的科学性和合理性。
Academicians have done many theoretical and practical researches of the Dempster-Shafer Evidence Theory. These researches have led to enriched and credible research findings, and these findings have contributed to the impetus functions of the evidence theory development. Because the relationship between the basic reliable numerical values sometimes cannot reflect the actual connection between the evidence, it can also lead to the ignorance of the correlation between economic and technical evidence. Therefore, relying only on the relationship between numerical values to modify the evidence can lead to erroneous results. Meanwhile, it can also decrease the interpretability and expansibility of results. In order to solve this problem, a new and concrete solution of the issue, which is based on evidence source feature, is proposed. In consideration of the classical evidence framework to which the second dimension information frame is added, and by using the second dimension framework, the evidence grouping criterion and evidence correction factor are confirmed. Different synthetic rules are used to compose the internal and external credibility of the evidence group. The second dimensional evidence framework includes not only the support of the evidence source for the focal element, but also the characteristics of the source itself, which can increase the information content of the framework. Above all, according to the issues of information fusion that analysts of security market have forecasted, the securities analysts＇ information concerning stock rating and earnings forecast are used to construct a two-dimensional progressive evidence framework. Using the method of evidence grouping to fuse the information, which is based on the two-dimensional recognition framework, this paper compares the results of different methods, including traditional methods of information fusion, classical Dumpster＇s synthesis method, and average evidence synthesis method. The proposed evidence grouping method can manage conflict information harmoniously. The method can also improve the accuracy of information fusion.
作者
朱卫东
王凯
吴勇
孙宜博
ZHU Wei-dong;WANG Kai;WU Yong;SUN Yi-bo(Hefei University of Technology,Hefei 230009,China;Key Laboratory of Process Optimization and Intelligent Decision-making,Ministry of Education,Hefei 230009,China)
出处
《管理工程学报》
CSSCI
CSCD
北大核心
2018年第3期119-125,共7页
Journal of Industrial Engineering and Engineering Management
基金
国家自然科学基金资助项目(71071048
71302063)
安徽省自然科学基金资助项目(1408085QG143)
安徽省软科学研究计划资助项目(1402052006)
关键词
D-S证据理论
两维证据框架
证据分组合成方法
证据源特征
Dempster-Shafer evidence theory
Two-dimensional frames of discernment
Grouping method for combining evidence
Characteristics of evidence
