Mining Periodic Patterns in an Attributed Graph Sequence

Due to the wide application of attributed graph sequence data in real life, it is increasingly urgent to mine the periodic patterns in attributed graph sequences. The existing algorithms can only mine the periodic patterns in transaction databases and graph sequences, which cannot meet the need of periodic pattern mining in attributed graph sequences. Therefore, we extend the task of periodic pattern mining in a graph sequence to an attributed graph sequence, and propose a method for mining periodic patterns in an attributed graph sequence. We firstly transform the attributed graph sequence into a non-attributed graph sequence based on the method of attribute-extended graph. Then a new algorithm is designed to mine periodic patterns in a non-attributed graph sequence. By mining the periodic patterns in the transformed non-attribute graph sequence, all the periodic patterns in the original attributed graph sequence are obtained. Experiments on real data sets validate the effectiveness of the proposed algorithm for periodic pattern mining.