Entropy-based Graph Clustering - A Simulated Annealing Approach

摘要

We revisit a Renyi entropy based measure introduced originally for image clustering [1], and study its application to graph clustering. To effectuate Renyi entropy based graph clustering, we propose a simulated annealing algorithm. We explore our algorithm's efficacy and limitations with the Karate club graph [2], as well as some other real world network.