Unified user malicious behavior detecting method and system in social network

Abstract

The invention discloses a unified user malicious behavior detecting method and system in a social network. The method comprises the steps of: establishing predicates and functions according to characteristics of multiple OSNs, collecting relation graph G and activity traces S, and constructing structural closed atoms and active closed atoms, then extracting the structural closed atoms, and mergingto form a training database; obtaining the first-order logic detecting and judging the malicious behavior; learn the weight of the first-order logic rule based on detecting results and the training database and establishing an MLN model; collecting a relationship diagram and activity traces of to-be-detected OSN and constructing structural closed atoms and active closed atoms, forming a to-be-detected database, and using the MLN model for detection. The unified user malicious behavior detecting method and system in the social network combines the OSN malicious behavior detection methods through the MLN model, realizes the joint detection of multiple malicious behaviors at the same time, and achieves higher detection accuracy. The unified user malicious behavior detecting method and systemin the social network can be widely applied in the field of information security.

Description

technical field [0001] The invention relates to the field of information security, in particular to a unified user malicious behavior detection method and system in a social network. Background technique [0002] In recent years, online social networks (OSNs for short) such as Facebook, Google+, Renren, Sina Weibo, etc., have been accepted and used by more and more people, accumulating a large number of users. But at the same time, OSNs are also subject to a large number of security and privacy threats, including privacy violations, spam, Sybil attacks, phishing, etc. These malicious behaviors violate the rights and interests of users and cause a lot of inconvenience to users and everyone and loss. Therefore, a malicious behavior detection system that can effectively protect users has important practical significance. [0003] Current malicious behavior detection can be broadly classified into structure-based and activity-based. Structure-based malicious behavior detectio...

AI Technical Summary

Problems solved by technology

[0004] Due to the prevalence and complexity of malicious behaviors in online social networks, each of the methods based on the above two models is only a specific malicious behavior detection for a specific social network, so the detection effect of different social networks and different malicious behaviors Not ideal; and because of the difficulty in combining the methods of the above two modes, it is difficult to improve the accuracy of detection by combining the above two modes

Images