Optimized Extreme Learning Clustering and Orthogonally Projected User Grouping for Online Social Networks

In social media, organizing friendship relationships is difficult since the more number of increasing users in Online Social Networks (OSN). To overcome this challenge, users in OSN heavily depend on grouping which is considered to be advantageous but at the same time found to be more cumbersome. More recently, recommender system and novel data clustering algorithm have been presented in OSN to address this concern. However, with the increasing nature and size of data (i.e., big data), grouping of users in OSN has become opened research for several academic’s professionals. In this paper, a novel clustering algorithm with optimized extreme machine learning which is called, Optimized Extreme Machine Learning and Orthogonally Projected (OEML-OP) is presented for user grouping in OSN to reducing computational overhead and time from large streams of social data. OEML-OP method utilizes new mechanism in EML to include optimality, a mechanism that is inspired by modularity function. The method performs user grouping through three main steps including, Duality Proportionality Graphical model, identifying optimal clusters and grouping of users in OSN. The Duality Proportionality Graphical model is to generate cluster of cliques and optimal clusters for the second steps and the Orthogonal Projection maximize the margin for separation between clusters. Due to the collective mechanism, the OEML-OP method gives better clustering accuracy and provides a novel model of grouping users on the basis of their activities. The analysis shows that the proposed method provides preferable clustering results and imparts a novel use-case of user grouping in OSN based on their activities.