Research On Social Bots Detection And Malicious Behavior Analysis

With the popularity of online social networks,the number of users in online social networks has grown exponentially.In order to provide users with a better experience and quality of service,social bots are increasingly being used to perform automated services.However,social bots are also used by malicious users to disseminate false information,which affect and deceive normal users in social network platforms,seriously jeopardizing user information security and environment of social network platforms.So detecting malicious social bots in online social networks has become especially important.In order to more accurately detect malicious intelligent accounts,such as malicious social bots,in social networks and protect user and platform information security.We propose a malicious social bots detection method based on clickstream sequence to detect malicious accounts in social network platform.At the same time,we conducts further analysis and research on the behavior of malicious social bots,which obtains the behavior sequence pattern of malicious social bots,and generates a behavior sequence pattern library.The main research contents and innovations of this topic are as follows:1.Based on clickstream sequences,a method of detecting malicious social bots is established.In social networking platforms,malicious social bots can more intelligently disguise by filling in personal information and other means,making it more difficult to detect.Based on the situation analytics theory,we obtain more robust user line dynamic features and time interval features in the time dimension by analyzing user clickstream data in depth.Then a semi-supervised clustering detection method based on transition probability features and time feature of clickstream sequences is proposed,which can timely and accurately detect malicious social robots in online social network platform.2.Propose a method of behavioral analysis and behavior pattern mining of malicious social bots based on the situation analytics theory.For the detected malicious social bots,we obtain the user's clickstream sequences set in the situation analytics environment,and mine the frequent behavior sequence pattern of the malicious social bots,then get the intention of the malicious social bots in the platform.Predicting the new behavior of malicious social bots can verify the proposed behavior patterns and discover new behavior sequences.Finally,the social bots behavior pattern library is built for the characteristics of the platform to achieve the purpose of longterm monitoring of malicious social bots.3.Application of malicious social bots detection technology in multimedia social network platform.Deploying malicious social bots detection method proposed by this paper in a multimedia social networking platform,namely CyVOD.It can detect existing malicious social bots accounts in the platform and evaluate the accuracy of social bots detection methods.Findings from our experiments on real online social network platforms demonstrate that detection accuracy for different types of malicious social bots by the detection method of malicious social bots based on transition probability of user behavior clickstreams increases by an average of 12.8%,in comparison to the detection method based on quantitative analysis of user behavior.The stability of online social network platform and the information security of users have become the concern of network users.In the social network environment,this topic detects and analyzes the social bot accounts existing in the social network platform.We propose a semi-supervised clustering detection method based on clickstream sequence transition probability feature and time feature for real-time detection of social bots in online social network platform,and then we establish a behavior sequence pattern library of malicious social bots to realize long-term monitoring of malicious social bots.