User Behavior Credibility Evaluation Model And Application Research Based On Intuitionistic Fuzzy Theory

With the increasing number of Internet users,the development of the Internet is changing rapidly.Faced with a surge of users and potential security threats,the traditional user behavior management system can no longer meet the needs of today's increasingly diverse,real-time and efficient development of network user platforms.Faced with an open,shared and complex network environment,the behavior of network users should be credibly evaluated,and real-time behavior data of users should be collected to effectively prevent the damage to the network system caused by the untrusted behavior of users.Because user behaviors are complex and diverse,the behavior itself is full of uncertainty,and the evaluation subject has a subjective evaluation willingness,so the difficulty of credible evaluation is how to quantitatively describe this uncertainty and subjectivity.Intuitionistic fuzzy theory can quantify the research object of the "other than the other" feature,and quantify the degree of uncertainty of user behavior through intuitionistic fuzzy number,and it is very suitable for applying it to the credible evaluation of user behavior.The thesis summarizes and analyzes the limitations of the existing evaluation model,combines the characteristics of user behavior uncertainty,introduces intuitionistic fuzzy theory,and proposes a credible evaluation model of user behavior combining AHP and intuitionistic fuzzy theory.Through simulation verification,the effectiveness of the model is further verified.On the basis of completing the evaluation model and combining the model itself as the essence of static evaluation,a dynamic evaluation mechanism based on sliding window and an incentive algorithm based on reward and punishment mechanism are proposed to optimize and expand the model.The dynamic evaluation mechanism starts from the actual user behavior and considers the impact of the user's behavior over time.From the perspective of trust management,the incentive algorithm takes reward and punishment measures for users who complete the trust evaluation,and effectively encourages users to restrict their own behaviors to achieve the purpose of maintaining network security.Specifically,the thesis mainly completed the following work:(1)A credible evaluation model of user behavior combining AHP and intuitionistic fuzzy theory is proposed.AHP realizes that behavioral evidence can still classify and stratify evidence better in complex situations.It is one of the classic models of multi-attribute decision-making problems.Intuitionistic fuzzy theory can realize the qualitative to quantitative transformation of the uncertain behavior and the subjective evaluation description in the process of behavior judgment,thus achieving the subjective and objective unity of behavior evidence and judgment methods.(2)A dynamic trust evaluation model based on sliding window and intuitionistic fuzzy entropy is proposed to realize the structural optimization and real-time update of behavioral evidence.Combining with the limitations of the existing sliding window mechanism,this paper makes reasonable improvements to the initialization of the sliding window.(3)An improved user behavior trust reward and punishment mechanism is proposed.The user behavior incentive algorithm based on reward and punishment mechanism realizes the feedback regulation of user behavior.The system can make corresponding reward and punishment measures according to the recent changes in the user's trust.After improving the reward and punishment mechanism,it is more in line with the characteristics of trust "fast decline and slow rise" and has good scalability.