Approximate Subgraph Mining Algorithm for Social Networks

Abstract

The application of graph mining is becoming more and more widespread, where approximate subgraph mining is one of the core techniques. However, the existing approximate subgraph mining algorithms have low computational efficiency and suffer from uneven subgraph identification. To address these problems, we propose an approximate mining algorithm ExMCMC-Motifs based on a Markov chain Monte Carlo sampling strategy with a common substructure. First, the vertices in the original network are sampled. Then the subgraphs involved in this vertex are identified using the MCMC random wandering sampling strategy,Finally, the neighbors of this vertex are sampled several times to achieve sampling equalization. The experimental results verify that the algorithm is computationally efficient and works well.