A new approach for finding communities of edges in complex networks

Discovering dense subparts, called communities, in complex networks is a fundamental issue in data analysis. A popular way to do this is to create a partition of the network. This partition can either be a partition of nodes, or a partition of edges. In this thesis I propose a new approach to finding a partition of the edges, by mimicking the approach of the Louvain algorithm, one of the most popular methods for node partitions. The Louvain algorithm is a greedy optimization technique using modularity as an objective function. I propose several different objective functions, edge modularities, to optimize in this approach and test the algorithm with different edge modularities on real networks