Community detection in networks via a spectral heuristic based on the clustering coefficient

The community detection problem in networks consists of determining a clustering of “related” vertices in a graph or network. Nowadays, studies involving this problem are primarily composed of modularity maximization based heuristics. In this paper, the author proposes a spectral heuristic based on a measure known as clustering coefficient to detect communities in networks. This measure favors clusterings with a strong neighborhood structure inside clusters, apparently, overcoming the scale deficiency of the modularity maximization problem. The computational experiments indicate a very successful performance by the proposed heuristic in comparison with other community detection heuristics in the literature.