一种集成区间逼近理想点的主观信任模型  

A Subjective Trust Model Based on Integrated Interval Similarity to Ideal

作　　者：徐军[1,2] 钟元生[1] 

机构地区：[1]江西财经大学信息管理学院,南昌330013 [2]江西财经大学现代经济管理学院,南昌330013

出　　处：《小型微型计算机系统》2014年第11期2428-2433,共6页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(71061006;61263018)资助;江西省自然科学基金项目(20114BAB201012;20132BAB201050)资助;江西省教育厅科技项目(GJJ12265;GJJ12740)资助

摘　　要：由于用户主观思维的模糊性和不确定性,电子商务环境下实体信任评价属性难以用精确数学模型表达,对此,提出一种集成区间的主观信任模型.模型利用集值统计理论对实体的某些关键属性用区间数表示,针对不同服务中评价属性权重难以客观确定的问题,提出基于熵权的属性权重确定方法,将属性集成区间序列与正负理想序列的相对接近度定义为实体的信任水平,利用逼近于理想区间灰关联系数的计算思想对信任水平进行排序.实例说明,实体的信任可以被量化,且实体间的信任具有可比性.最后,通过实验仿真,证明该方法在描述实体平均信任水平及抗欺诈方面具有良好的效果.For fuzziness and uncertainty of users subjective thinking, entity trust evaluation attributes are difficult to be described accu- rately with mathematical models. In order to solve this problem, we propose a subjective trust model based on integrated interval. This model measures the entity＇s some key attributes by using set-valued statistics theory, defines the entity＇s trust level as relative closeness of its list and references in positive and negative ideal gray relational coefficient matrix. Furthermore, a computation method based on similarity to ideal interval gray relational coefficient, which combines the grey relation degree, entropy weight and TOPSIS, and is used to order trust level. In the example, it is explained the entity＇s trust is quantified and trust among entities can be compared. Finally, through simulation, it proves that this method can obtain good effect in expressing entities average trust level and Anti-fraud.

关 键 词：集成区间 主观信任模型 熵权 灰关联系数 理想点法 相对接近度 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]