A Hybrid Contextual Probabilistic Graph Clustering and Link Prediction Model for Complex Social Networking Data

Abstract

Probabilistic based graph community detection plays a vital role in the complex social networking datasets. Since, most of the conventional approaches are difficult to predict the new type of link prediction using the standard graph community clustering measures. Also, traditional clustering measures use nearest neighbour measures instead of contextual similarity in order to predict the relationship among the different graph nodes. In order to optimize the contextual node clustering and link prediction, a hybrid dynamic scalable measure is proposed for the community clustering on complex networks. In this work, a hybrid graph clustering and link prediction approaches are proposed on the complex social networking dataset for better decision making patterns. Experimental results prove that the proposed contextual probabilistic graph clustering and link prediction approach has better efficiency than then conventional models on complex social networking datasets.