Extraction of community transition rules from social bookmark data as graph sequence

In this study, we use transaction data collected from a data stream regarded as a graph representing the change in structure and sequence data for each relevant time period to analyze the changes in the sequence graph of a community. The algorithm proposed in this paper uses the hierarchical clustering method combined with a graph kernel extension to analyze the relationship among the chart series of the entire community. Extracted community rules appear occasionally, and then are shown to disappear in the middle of the series. The results of experiments using synthetic datasets and real social bookmark data show that changes in the community captured occasional occurrence of the proposed algorithm.