Understanding and Predicting Private Interactions in Underground Forums

摘要: The studies on underground forums and marketplaces have significantly advanced our understandings of cybercrime workflows economies. Researchers economies conducted comprehensive public interactions. However, little research focuses private lack the investigation interactions may cause misunderstandings economies, as users in tend to share minimal amount information resort messages for follow-up conversations. In this paper, we propose methods investigate analyze a recently leaked dataset from Nulled.io. We present analyses contents purposes messages. addition, design machine learning-based models that only use publicly available detect if two privately communicate with each other. Finally, perform adversarial analysis evaluate robustness detector different types attacks.