Privacy Protected System for Vulnerable Users and Cloning Profile Detection Using Data Mining Approaches

Abstract

The evolution of Online Social Networks in recent years have lead to an enormous rise in users working on Facebook, LinkedIn and twitter etc.. Actively, it can be claimed that Internet projects itself as the ultimate mode of communication providing easy and user friendly exchange of information worldwide. The figures reveal that OSN namely Facebook, LinkedIn and twitter etc have minimum one profile among the 80% of active online users. Such users exchange and post information or write ups based on every days life related topics that vary from events, politics, news and celebrities. As far as OSN is concerned this kind of behavioral outlook is seriously refrained by the privacy settings concerning public profile data. These social network offers restrictive data related to user profile that is publically available. These networks face a big security challenge which is also a prime concern for the users too and that is the building up of fake accounts and hacking users personal data. The research paper suggests a Privacy Protected System technique which identifies malicious user and examines any sort of attack thus enhancing the security and making the network protected. Privacy protected System is categorized under 2 types: One is the vulnerable user protecting which detects malicious user while other being the Profile cloning identification and detection. The main motive of suggested Privacy Protected Systems being helping in vulnerable user detection yielding in greater detection rate on available intrusions along with the feature of detectors in detecting malicious attacks. The experiment reveals that by implementing our technique there was noticeable outcome in identifying unknown attacks that was higher than with detection rate as 97.29 % and below 1% of false alarm rate.