Research On Trust Evaluation Mechanism Based On Improved D-S Evidence Theory

In recent years, with the rise and development of the Internet, the existing trust evaluation mechanism hassome shortcomings, such as not handled quickly and efficiently evaluate the mechanism for malicious attacks, cannot meet the practical requirements of the nodes on the assessment of the accuracy can not effectively identify andprevention of collusion node attacks. To solve the above problems, this paper trust evaluation mechanism made athorough research and made a credible evaluation mechanism based on D-S evidence theory. Specific work madeas follows:A trust evaluation model based on improved D-S evidence theory and its implementation framework areproposed firstly. It expresses a trust with a triple based on D-S evidence theory to take the characteristics of trustpossibility into account. It improves basic probability function, use a sequence factor to find entity’s continuousoscillation attacks and conduct spoofing attacks. It gives a specific algorithm to find continuous sequence to makethe model more practical. It triples relationship of trust based on D-S evidence theory and proposes a trust-basedevaluation function normalization method to measure trust more currectly. Experiments show that the model cansuppress faster monomer malicious behavior, and make assessments closer to the actual value.In Section4, the proposed model is a collusion detector based on G-N algorithm for trust model. There is atouted feature of collsion networking group members by combing sociological relationships and ananlysing theactivity to make this detector model useful in bad collusion attacks. It gives a specific algorithm to calculate thecoefficient of faltting to make model more practical. It designs a divided algorithm based on G-N clusteringalgorithm to reduce the datasets alternatives of suspicious community to make the efficiency of detector modelhigher. It identifies collusion malicious group by assessing the damaging impact and the degree of collusion, sothat the trust model can better suppress conspiracy aggressive behavior. Experimental results show that the modelhas high recognition accuracy in identifying aspects of conspiracy nodes, enhanced security and reliability of trustevaluation mechanism.Finally, we design and implement a simulation system to analyse feasibility and validity of this trustevaluation mechanism.