Colluding Identity Clone Prediction using Machine Learning

Abstract: As Online Social Networks such as Facebook, Linkedln and Twitter are increasing becoming a part and parcel of one's daily lives, personal information is at stake. Easy access to personal information has made the attackers  to  steal  information  from influential  users using various forms of attacks. Attackers take advantage of the user’s trustworthiness when using Online Social Networks. Hence, there is a need for the third party applications of various Online Social Networks sites to provide defence mechanisms against adversaries. Colluding  attack is a way of creating  fake  profiles  of  friends  of  the target in the same OSN or others. Colluders impersonate their victims and send friend requests to the target with an intention  to  infiltrate  their  private  circle  to  steal information. These types of  attacks are difficult to detect in  because  multiple  malicious  users  may have  a  similar  purpose  to  gain  information  from their targeted user. In this regard, the work intends to overcome  this  type  of  attack  by  addressing  the problem of  identity clones across multiple Online Social Networks  using machine learning.

Keywords: Identity Clone Attack, Machine Learning,  Predictive FP growth, Online Social Networks