Study Of Web User Behavior Trust Evaluation And Behavior Control Methods

With the increasing development of Web technology,It provides rich technology of front end and strong background support to promote the generation of new business model,such as e-commerce and e-government.But when web user’s account be stolen by Trojan or sustain information leakage that caused by suspicious links,It will cause the identity of the web user cannot be trusted.Or the user’s identity is credible,but the user’s behavior is not be trusted when people operate the system,How to ensure the security of web user’s information becomes the urgent task.There are a large number behavior of user in web application practice,some of these behavior represent the user’s personal habits,By studying the web user’s operating habits,we can identify the user’s abnormal behavior that does not conform user’s habits,These abnormal are caused by the security problems of user’s account.On the other hand,when we design behavior identify algorithm,due to the trust evaluation of user behavior is based on a large amount of evidence,So simply rely on subjective weighting method will lead to the result is too subjective,can’t get all kinds of objective changes that belong to evidence,the second,user behavior is a series actions for a long time,only evaluating one of the user’s behavior is mistrust,Putting forward a model also need to consider the ease of use.From what has been discussed above,the trust evaluation of user need a variety of user behaviors as a basis,in the condition of subjective and objective Angle have a relative balance,combined with science and reasonable evaluate method to get the user behavior’s trust evaluating value,through evaluating value,we can control the user’s access to meet the safety demand of network platform.In this article,through building electric business platform,on the basic of a large number of user’s behavior evidence,proposing a user behavior evaluating model that organic combine entropy weight method with AHP analytic hierarchy process(AHP),This model is a dynamic model of sliding window,when we evaluating the user behavior,each evaluation will combine with the latest user behavior and historical evidence to get a new set of dimensionless evidence,And then using the entropy weight method and the dimensionless evidence to get the objective weight,through AHP analytic hierarchyprocess(AHP) obtain subjective weight of the behavior evidence, finally,taking a optimization processing way to deal with subjective weight and objective weight,we can getting a integrated weight that conform to the actual requirements.Through the practical applications and theoretical analysis,The model allows the subjective and objective have a relative balance,can effectively distinguish behavioral types of trusted user such as dangerous user and malicious user with a lower false positives ratio,laying the foundation of key network application’ s safety regulation in theory and practice,the key network applications include e-commerce and online financial and so on.