Sampling Subgraph Network With Application to Graph Classification

Graphs are naturally used to describe the structures of various real-world systems in biology, society, computer science etc., where subgraphs or motifs as basic blocks play an important role function expression and information processing. However, existing research focuses on statistics certain motifs, largely ignoring connection patterns among them. Recently, a subgraph network (SGN) model is proposed study potential structure it was found that integration SGN can enhance series graph classification methods. lacks diversity quite high time complexity, making difficult widely apply practice. In this paper, we introduce sampling strategies into SGN, design novel model, which scale-controllable higher diversity. We also present structural feature fusion framework integrate features diverse SGNs, so improve performance classification. Extensive experiments demonstrate that, by comparing with our new indeed has much lower complexity (reduced two orders magnitude)