DETECTION OF COMPROMISED ACCOUNTS IN ONLINE SOCIAL NETWORK

Compromised accounts are of a severe risk to the social network users. People now a days are mostly dependent on Online Social Networks. While some persistent spams feat the relationship between the users by spreading spams. Therefore time to time detection of the compromised accounts is a necessity. In this paper, we will study different social user behaviour and detect the compromised accounts and spam users. Spam behaviour in social networks has a wide range of illegal activities. Such activities need to be evaluated and effect of spam users needs to be reduced. To reduce such effects, we require proper detection strategy. We validate the effectiveness of these behaviour by collecting the clickstream data on a social network website. Social behaviour reflects the users behaviour online. While an legitimate user coordinates its social behaviour carefully, it is hard for the fake users to pretend to be affected. Different studies are performed in spam behaviour analysis and define a structure for spam account detection.

I. Introduction

We live in the generation of social networks. Like online social networks facebook and twitter have become quite popular in couple of years. Many people use social networks to perform different activities online like staying in contact with relatives, friends, chatting and also acquiring information and news. This creates interactions between different users forming an activity log. These information are stored by the social network service providers. The access of such information to a spam user may expoit the privacy and security of the social network. It may also spread some illegal content and wrong information. These users or accounts gain access to information of other users or spread unwanted messages. Comman spam behaviors like Sybil attacks occur in online social networks. Therefore identifying the spam user is a necessity.