Benchmark snapshots for testing social network evolving algorithms.

A new evolving network model with community structures is proposed and analyzed. The formation of communities is driven by each node’s various preference of interest fields, denoted as interest vectors, whose mechanism is far different from other models with community structures. The theoretical analysis and numerical simulation show that the degree distribution of the model represents a transition between power-law and exponential, depending on parameters given, which is close to real-life networks. The model can generate a series of snapshots of the networks, varying with time, with well-defined communities, which could be taken as benchmark graphs for testing community detection algorithms. Because the community structure of each time slice proves to be reasonable, it can also reasonably represent the evolution of the community.