Study On Evaluation Of User Behavior Trust In Trusted Network

With the rapid development of network technology and newly additional applications, network security is confronting many challenges. The fundamental reason is the weakness of network system, so studies of trusted network becomes a new trend in network research and development. Based on present security technology and starting with user security, the network trusted technology could resolve the trusted security problems in the entire network by adding behavior trust and establishing trusted connections between users in different networks. Traditional security mechanism could resolve the problem of user identity trust, which could not ensure the user behavior trust. By establishing an evaluation model of user behavior trust, predicting and controlling the user behavior with evaluation results, the user behavior trust could be implemented, which becomes the essential content of trusted network research.In this thesis, the author has analyzed characteristics of the network user behavior trust, confirmed basic principles of the evaluation of user behavior trust, on which basis reviewed decision attributes that affects the user behavior trust, established evaluation model MDAM of user behavior trust, and evaluated the user behavior trust by applying quantitative evaluation based on multiple decision attributes. Some main tasks have been achieved in the following:①Measure methods of user behavior trust attributes, selecting methods of evaluation indexes system, the pre-processing of evaluation indexes and weight determination methods are researched. A comparison is made between present evaluation models of user behavior trust, which provides theoretical basis for establishing new evaluation model of user behavior trust.②An evaluation model MDAM of user behavior trust based on multiple decision attributes is put forward, which implements the evaluation of user behavior trust degree. User behavior evidences are data basis of MDAM modeling, and methods of acquiring user behavior evidences through MDAM and standardization process are discussed. On this basis, through evaluating the user behavior trust degree with multiple decisions attributes like direct trust degree, indirect trust degree, trust risk function, behavior active degree, and behavior reward and punishment factor, computational algorithm of each decision attributes, objective weighting method based on information entropy, and integrating method of decision attributes about weighted geometric mean are researched. The MDAM introduces risk mechanism, reward and punishment mechanism and behavior active degree to improve accuracy and security of the evaluation, and through an establishment of evaluation degree space of the user behavior trust to form a mapping mechanism of user behavior degree and service degree, the user behavior trusted management is implemented by MDAM.③Through simulation experiment, the process of accomplishing the evaluation of user behavior trust with MDAM is presented, and results of modeling evaluation are analyzed. MDAM has computed the direct trust degree with Yaahp software based on AHP and other decision attributes trust degree and weight, gets overall trust degree through weighted geometric mean method, and lastly implemented user behavior trusted management according to evaluation space and mapping mechanism. Performance self-examination of the MDAM is put up from aspects of evaluation accuracy and security with Matlab software, which is took as performance evaluation tools, and comparison is made between PET and EigenRep models, which verifies the result of MDAM user behavior evaluation.