CP-GNN: A Software for Community Detection in Heterogeneous Information Networks[Formula presented]

Recently, the topic of community detection (CD) in heterogeneous information networks (HINs), which contain multiple types of nodes and edges, has received much attention. However, existing CD methods could not well exploit the high-order relationship among the nodes to detect communities. To alleviate this issue, we propose to use the concept of context path to model the high-order relationship among nodes, and develop a novel context path-based graph neural network (GNN) software, called CP-GNN. It can not only learn the embeddings of nodes for detecting communities accurately, but also well capture the high-order relations between nodes unsupervisedly, which enables great convenience to the study of communities and real-world applications.