Profile Similarity Technique for Detection of Duplicate Profiles in Online Social Network , Dr. M. Nandhini 1 , Bikram Bikash Das 2 1 Assistant Professor, Dept of computer Science Pondicherry University, India 2 Department of computer science, Pondicherry University Kalapet, R V Nagar, India Abstract— In the current generation social network has become a popular way to communicate with each other which are spread across diverse location around the world. In social network any user can find other users and make friendship and they can make friend circle online around the world and thus users can form their own network. An Individual user can have multiple social network accounts to keep in touch with friends in many social networking sites. Online Social Network users are not aware of the various security attacks like privacy violation, identity theft etc. Any user can create fake profiles with the name of real user. Other online social users will think it as real users and they might be responded to them which are not actually the real user. It makes the whole network quite confusing and frustrating. In this paper, we will provide a similarity technique which can analyze social network data based on attributes similarity. The proposed system can detect as many similar social network profiles as possible and analyse them in order to find whether it belongs to same or different persons. It makes other user easy to communicate with each other in a safe and efficient manner. Keywords—Social Network Analysis, Social Engineering Attack; Duplicate profiles; Global profile Database, Profile attributes matching, Suspicious profiles I. INTRODUCTION Social network makes our digital life become simple to have social relationship with friends online and social network websites such as Facebook, MySpace, Google+ for connecting peoples, YouTube for video sharing, Google+, LinkedIn for professional identity, Tweeter for updating daily tweets (messages) of some event etc. are becoming popular social network website used among peoples of all ages especially among youths. These social network sites are famous among internet users and users are interconnected to each other via online social relationship known as friendship. Various social network sites has been developed to gain their attraction among people where any use can get can membership by fill up a simple registration form. An individual user can make multiple accounts with his same attributes such as E mail id or mobile number in many social networks. [1] An existing user can have many numbers of social network profiles within the same or different network. It makes social network vulnerable to attack by using someone similar attributes. Most of the cyber crimes are happening in social network sites. [2] Any user can make profiles with others attributes such a same name, college, age, profile images etc. Fake profiles are being created in all the social sites and victim personal information is becoming more and more vulnerable to attack by the attacker in various ways. In many cases Name can be same for many user but the other attributes such as profile image, qualification, address and mobile numbers all can’t be same for multiple social network users[1][2]. The main idea behind the proposed approach is to find out as many social network users which have similar attributes and to find out the originality or real users from it. In recent Research reveals that almost 80% of profiles in face book are fake one. Any user can make some fake accounts by using others attributes to fool other users. [3] Many users’ like to disclose their personal information like phone no., date of birth, address etc in their profiles. Availability and revealing of such personal information might be the sources of profile data that the attacker is trying to get access to create similar profiles. Any other user can create Fake profiles in the name of the real user with that personal information and try to launch various attacks such as sending and posting irreverent messages and tries to fool others to get the confidential information.[4] The proposed approach can detect as many similar social network profiles as possible based on the similarity of profile attributes and analyse them in order to find whether it belongs to same or different persons. The publicly available profiles and their attributes are extracted and then store them in global database of profiles so that their existence can be checked in many social networks. It helps the others user to have online social communication with each other in a safe and efficient manner. II. RELATED WORKS Many of cybercrimes are because of the facilities to create unlimited numbers of profiles by the same person within the same or different networks who try to act as real users and violating the rules. It makes other users difficult to identify who are the real users and who are fake. Thus social network analysis comes to the scenario which is a new area of research that involves analysing the network structure for the benefit of society so that online user can have social connection and conversation in a safe way with other user around the globe. In recent scenarios it becomes a serious problem and many researchers have begun their research in identifying the real user’s identity. Sophia Alim, Daniel Neagu ,Ruqayya ISSN:0975-9646 M. Nandhini et al, / (IJCSIT) International Journal of Computer Science and Information Technologies, Vol. 7 (2) , 2016, 507-512 www.ijcsit.com 507