Structual Change Point Detection for Evolutional Networks

We propose a change point detection algorithm for a sequence of graphs. Our algorithm focuses on the change of the structure of densely connected subgraphs (community structure) rather than the change of the link weights. In contrast to the traditional approaches, the algorithm can identify the structure change more sensitively. Experiments with a synthetic data and a real-world data of graphs showed that our algorithm can accurately locate the changed subgraph compared with some of the state-of-the-art algorithms