Machine learning techniques for spammer identification: State of the art and analysis

Abstract

Internet offers seamless communication with a high consumption rate. With the increase in the widespread utilization of internet tools such as e-mails, the spammers started to exploit the e-mail network to effect malicious and hazardous activities. As the e-mail spam detection systems became more sophisticated and accurate, the attention of spammers has now turned to the recently emerged social networks due to massive usage and popularity among the people. A range of online social networks (OSN) such as Facebook, Twitter, Instagram are available with each of them offering unique services to the account owners and benefits both professionally and personally. Besides these advantages, the network also houses illicit accounts that disturbs the user experience and destroy the ultimate objective of social networking. The existing techniques employed by these OSN to detect such illicit users are not effective and accurate and also demands manual approaches to spot them. Hence, a number of methods and algorithms were proposed in the literature to identify the spammers concealed in these networks. This work attempts to provide a detailed review of state of the art techniques and methodologies employed in the spam account detection problem along with future research directions.