User profile matching and identification using TLBO and clustering approach over social networks

Abstract

User existence increases a lot in recent time over online social networking sites. Each social networking site has its importance and purposes, which leads to having duplicate accounts of the same user over different sites and poses redundancy of information over the Internet. Literature has proposed identity search methods by profile, content, and network attribute but these methods are time-consuming. This paper explores the problem of user profile duplication and proposes a time efficient profile matching algorithm across two networks to identify the similar user. Authors studied a well-known soft computing method of clustering for identifying the duplicate users and propose a hybrid method using Teacher-Learner Based Optimization (TLBO) with clustering to achieve a higher rate of accuracy and time efficiency. The experiment was carried out over real-time dataset generated using the facepager tool over Facebook and Twitter social Networking sites. Outcome result shows that proposed hybrid TLBO based clustering is showing better performance over different evaluation parameters.