Application of Machine Learning Method in Detecting the Spammers in Social Networks

Social networks have become one of the popular ways for users to meet and interact online. Users spend a lot of time on popular social networks (such as Facebook or Twitter). It is possible to store and share personal information, as well as the possibility of contacting thousands of users. This action can cause many problems. For example, cybercriminals may create trusting relationships between users in order to deceive them. As another example, cybercriminals may target valuable personal information for identity theft or to find directions to direct a spam campaign. In this article, we analyze how far spam has entered social networks. More specifically, we will analyze how spam enters, targeting social networking sites. A large and diverse set of "favorite profiles" were created on three major social networking sites to collect information about spam activity, and the types of calls and messages they received. We then analyze and identify the collected information and measure unusual user behavior. Based on the analysis of this behavior, we create methods to identify spam in social networks. Our results show that it is possible to automatically identify accounts used by spam, and with effort, this can be analyzed in a real-world social network. More specifically, this study was conducted on Twitter, which resulted in the removal of ۱۵,۸۵۷ spam profiles.

کلیدواژه ها:

Machine learning ، Spammers ، social networks