Local Community Detection Based on Core Nodes using Deep Feature Fusion

LI Xiao Hong

Abstract:

Unlike global community detection, local community detection is to identify a cluster of nodes sharing similar feature information based on a given seed. The accuracy of many local community detection algorithms heavily relies on the quality of seed nodes. Only high-quality seed nodes can accurately detect local communities. At the same time, the inability to effectively obtain node attributes and structural information also leads to an increase in subgraph clustering error rates. This paper proposes a Local Community Detection based on Core Nodes using deep feature fusion, named LCDCN. We find the nearest nodes for the seed nodes, then construct a k-subgraph through a specific subgraph extractor based on the core nodes. Subsequently, two deep encoders are employed to encode and fuse the attribute and structure information of the subgraph, respectively. Finally, the local community is discovered by optimizing the fused feature representation through a
self-supervised optimization function. Extensive experiments on 10 real and 4 synthetic datasets demonstrate that LCDCN outperforms its competitors in performance.