Abstract: Online social networks (OSNís) such as facebook, twitter, linked in, weibo have become important aspect of users in daily life. Social network have changed the way people interact, it fulfils user needs by information sharing and appreciation. OSNs suffer from fake profiles creation. An OSN user faces security problems like identity theft, privacy violation, leakage of personal information, identity theft. Fake users may inject spam, modify the online ratings and extract knowledge of genuine user. It is very difficult to detect, verify fake profiles manual. There is a need of automation of this fake profile userís identification process. This paper explores different machine learning techniques which can be used for classification of fake profiles and real users.

Keywords: online social networks, fake profiles, classification, machine learning algorithms.