A Weight-Based Approach: Frequent Graph Pattern Mining with Length-Decreasing Support Constraints Using Weighted Smallest Valid Extension

As one of the recent interesting areas in data mining, frequent graph pattern (or sub-graphs) mining was proposed to deal with more complicated data such as graphs. Moreover, in order to extract more meaningful graph patterns, the concept of weighted frequent graph pattern mining was suggested. Meanwhile, graph patterns extracted from given graph databases can have various features, which can be different for each length of the patterns. However, traditional graph mining approaches cannot deal with such useful characteristics in their mining operations because they employ a single minimum support threshold. Hence, even if a mined graph pattern turns out to be weighted infrequent, it may become a useful result according to its length feature. In this paper, we propose a new method for mining weighted frequent graph patterns based on length-decreasing support constraints and weighted smallest valid extension. Experimental results of this paper show that the proposed algorithm outperforms a state-of-the-art approach in various aspects.