Zhang, Yiming, et al. "Key player identification in underground forums over attributed heterogeneous information network embedding framework." Proceedings of the 28th ACM international conference on information and knowledge management. 2019.

The paper presents iDetective, an intelligent system designed to automate the analysis of underground forums for identifying key players in the cybercriminal ecosystem. iDetective uses an attributed heterogeneous information network (AHIN) and a meta-path based approach to represent users and capture semantic relationships. The system introduces Player2Vec, which maps AHIN to a multi-view network and applies graph convolutional networks (GCNs) and an attention mechanism to learn and fuse user embeddings for key player identification. Experiments on data from Hack Forums and Nulled validate iDetective’s effectiveness compared to alternative methods.