Modeling of social network services for deception detection

Social Networking Services (SNS) has become an important element of people’s daily lives. Users of SNS tend to send numerous messages and keep updating each other during engagement for information sharing. Meanwhile, however on the dark side, advertisers and malicious users are also attracted to it. The problem is getting more serious as the number of sites increases. Spam messages from advertisers and malicious users have been reported greatly annoying normal users. Accordingly, technologies referring to solve this problem is worth to be investigated.||Our study mainly focuses on the second largest Social Network Site in the world, Twitter. In this paper, the interaction methods and types of relations would be reviewed first. Consequently, several features of deception behavior are identified. These features are then computerized and tested against real data. Results are analyzed and statistics are generated to reflect whether the features identified earlier effectively represent certain aspects of spammers’ behavior prevailing in recent time. Our investigation provides some hints on detecting spam on Twitter, and it is hoped that the outcomes of the study can provide rooms for improvement of anti-spam systems in future development.