A method of user behavior credibility detection based on fuzzy entropy weight method and cloud model

Abstract

The invention belongs to the technical field of network data processing and discloses a method for detecting the credibility of user behavior based on a fuzzy entropy weight method and a cloud model,wherein, a behavior attribute cloud is established, a hierarchical cloud is established, membership degrees of m behavior clouds to n hierarchical clouds are calculated by an association formula, anda membership degree matrix is obtained accordingly. According to the fuzzy entropy weight method, the weight of each attribute of the membership matrix element is obtained. The evaluation system provided by the invention takes into account that the calculation is only around three digital characteristics of the cloud, does not involve more complex parameters, and the calculation process is simple,so that the final evaluation result is obtained by multiplying the membership degree matrix and the weight vector. The invention determines the evaluation matrix according to the relationship betweenthe attribute cloud and the hierarchical cloud, and the cloud model itself is a reflection of the uncertain nature of things, and corresponds to the randomness and uncertainty of behaviors, and the evaluation process does not involve subjective factors, and the evaluation result is more reasonable and credible.

Description

technical field [0001] The invention belongs to the technical field of network data processing, and in particular relates to a user behavior credibility detection method based on a fuzzy entropy weight method and a cloud model. Background technique [0002] At present, the existing technologies commonly used in the industry are as follows: [0003] Traditionally, the management of users is mainly aimed at the level of user identity information, including digital certificates, ID cards, biometrics, and so on. Research and analysis of Internet user behavior is relatively rare. Common methods include user behavior credibility evaluation models based on fuzzy theory, but when constructing weight vectors, the methods used are usually through expert experience or AHP. On the one hand, expert analysis is used to construct weight vectors, making the evaluation process The degree of supervisor participation is high, and the evaluation results are not objective enough. On the other ...

AI Technical Summary

Problems solved by technology

[0005] Although the traditional trust model can solve the problem of user identity authentication, after the user successfully accesses the network, a series of operations performed by the user are not monitored in real time, which allows malicious users to perform various tasks after forging a normal identity to access the network. Various attacks put the network at risk

[0006] Existing user behavior evaluation methods, such as tomographic analysis model, BP neural network analysis model, fuzzy theoretical analysis method, etc., ignore the interaction, interdependence and feedback relationship between elements of user behavior at various levels, and the evaluation process depends on The professional knowledge of experts and the degree of cognition of the evaluation system, as well as the subjective weight distribution, lead to unreasonable evaluation results and indirectly affect the establishment of a credible and controllable healthy network environment

[0007] These evaluation models and methods in the prior art do not consider the randomness and ambiguity of user behavior itself, and fail to take the uncertainty of behavior as the main factor affecting the evaluation results. The understanding of how user behavior affects network security not deep enough

Images