Identification of Malicious Accounts in Facebook

Abstract

In this modern world of internet age social networking is becoming popular day by day for sharing of information and day to day communication. The social media has changed the way people pursue their life and made it a part of their life to get connected with friends, family, colleagues and with the society as a whole. Social networking can occur for social purposes, business purposes by the use of Facebook, Twitter, LinkedIn etc. Hence it has become the favorite spot for the cyber criminals. Facebook is a free social networking website that allows users to upload photos and video, send messages and keep in touch with friends, family and rest of the world. Unfortunately, spammers are also using this Facebook for their personal gain by creating fake profiles or comprising the other famous accounts. These accounts together are referred as malicious accounts. Hence it became important task to identify this spam for free communication over Facebook. However, some former researchers have proposed their work to identify these accounts but suffering from many limitations. In this paper a method is proposed called An unsupervised method to detect spam in Facebook using DBSCAN algorithm. This research work used Jaccard similarity and DBSCAN algorithm to detect spam messages. This research work is implemented using real time messages and evaluated. The experimental results show the methodology used is very efficient, having better accuracy and low false positive rate in detecting the spam messages in Facebook. This research work will help the common man in detecting malicious Facebook accounts in a great way.