Finding communities is one of the most outstanding issues and of considerable interest in the analysis of complex networks. As a matter of facts many networks of scientific, social and biological interest are found to dividenaturally into communities.
Among the different approaches adopted for this task, one highly effective method is the optimization of a quantity called modularity. Here we present the implementation and the performance analysis of Newman’s modularity algorithm for communities detection and we discuss about features and limits of this approach.
We show the results of the algorithm both on toy-models, specifically implemented for this task, and real-world networks, taken from the Stanford Large Network Dataset Collection (SNAP).
For the project final report click here.
