원문정보
초록
영어
Trust evaluation is increasingly important for collaboration in the Internet today. We propose a multi-dimensional evidence-Based trust evaluation model named EBTrust model, which expands the types and sources of evidence. We put both transaction feedback and the network operating behavior information into this model. And we improve the D-S evidence theory and propose a new rule of evidence synthesis that based on the conflict intensity (G) and efficient conflict (Gh), so we call it G-Gh synthesis rule, which can not only solve the high confliction and complete conflict feature of multi-dimensional evidences but also the problem of uncertainty in the process of trust evaluation. The example analysis and experimental results show that the EBTrust model can resist conspiracy and malicious comment behavior.
목차
1. Introduction
2. Framework of EBTrust Model
3. Evidence Preprocessing
3.1 Preprocessing of E-commerce Feedback Evidence
3.2 Preprocessing of Online Community Feedback Evidence
3.3 Preprocessing of Operating Behavior Feedback Evidence
4. Construction of Basic Trust Function
5. G-Gh Synthesis Algorithm
5.1 Conflict intensity G
5.2 Weighted Average the Value of the Trust Allocation
5.3 Efficient Conflict Gh
5.4 Algorithm of G-Gh
6. Experiment and Analysis
6.1 Resistance of Conspiracy Behavior
6.2 Resistance of Malicious Comment Behavior
7. Conclusions
References