A General Optimization Technique for High Quality Community Detection in Complex Networks

Recent years have witnessed the development of a large body of algorithms for community detection in complex networks. Most of them are based upon the optimization of objective functions, among which modularity is the most common, though a number of alternatives have been suggested in the scientific literature. We present here an effective general search strategy for the optimization of various objective functions for community detection purposes. When applied to modularity, on both real-world and synthetic networks, our search strategy substantially outperforms the best existing algorithms in terms of final scores of the objective function; for description length, its performance is on par with the original Infomap algorithm. The execution time of our algorithm is on par with non-greedy alternatives present in literature, and networks of up to 10,000 nodes can be analyzed in time spans ranging from minutes to a few hours on average workstations, making our approach readily applicable to tasks which require the quality of partitioning to be as high as possible, and are not limited by strict time constraints. Finally, based on the most effective of the available optimization techniques, we compare the performance of modularity and code length as objective functions, in terms of the quality of the partitions one can achieve by optimizing them. To this end, we evaluated the ability of each objective function to reconstruct the underlying structure of a large set of synthetic and real-world networks.