Probabilistic Community Detection in Social Networks

The detection of community structures is a very crucial research area. problem has received considerable attention from large portion the scientific community. More importantly, these articles are spread across number different disciplines, computer science, to statistics, and social sciences. analysis modern networks becomes rather cumbersome, as their size keeps growing larger larger. Moreover, in communities, users participate groups. From network perspective, efficient methods should be developed automatically identify overlapping that is, communities with nodes. In this work, we use probabilistic model characterize linked common innovative idea work represented Markovian continuously changing states. Each state represents within cluster, have specific characteristic classes. Based on current state, introduce fast, linear newly added users, approach estimate probability each cluster homogeneous terms sets user characteristics determine how well new fit Because computations involved, our proposed can detect overlaps low execution time high accuracy, shown experimental results. results scheme executes faster provides more robust compared competitive schemes.