基于回归分析和节点互助的MRA-SMAN信任模型  

Trust Management Model Based on Analysis of Regression and Set of Mutual Assistance Nodes

作　　者：代致永 郜燕[1,2] 刘文芬[1,2] 薛长松[3] 

机构地区：[1]信息工程大学 [2]数学工程与先进计算国家重点实验室 [3]95879部队

出　　处：《信息工程大学学报》2014年第4期466-471,共6页Journal of Information Engineering University

基　　金：国家937计划资助项目(2012CB315905)

摘　　要：针对现有动态信任模型在不同网络环境下适应能力不足、信任信息利用不充分和健壮性不足的问题,文章提出了基于回归分析和节点互助的MRA-SMAN信任模型。通过建立回归分析子模型,使得模型能够满足各种网络环境下信任评估的需求,而模型中满意度的多因子表示方法的引入解决了已有模型中信任信息无法充分利用的问题。最后,通过建立SMAN节点互助集合,增加了所提模型对抗虚假信息的能力。仿真实验表明,模型在抵御策略型攻击和协同作弊攻击的过程中,体现了良好的收敛性、准确性以及适应性。Different network environment has different evaluation criteria about trust. It is also the reason why so many trust models lack adaptability. And the way to fully use trust information is to divide it into several details, but few models are aware of this problem. This paper presents a general trust model, called MRA-SMAN for a more comprehensive and accurate trust evaluation. It intro- duces a multi-factor analysis method to realize the ability to suit for several network environments, and uses multiple linear regression to calculate the weight of each factor, it also creates a set of mutual assistance nodes to solve the problem about recommendation trust. Simulation results show that the presented model reflects the better accuracy, convergence and adaptability in dealing with complex malicious attacks and fraud.

关 键 词：信任评估模型 多因子决策 多元回归分析 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]