Graph Neural Network and Natural Language Processing Integrated Methodology for Unmasking Scam Propagation in Social Network

The proliferation of scammers on social networks poses a significant threat to online security. This paper introduces a comprehensive methodology that integrates Natural Language Processing (NLP) and Graph Neural Networks (GNNs) to identify scammers on social networks based on both content and network connections. NLP algorithms identify suspicious words and phrases frequently used by scammers, while machine learning methods analyze phishing links. Additionally, GNNs and community detection algorithms expose networks of scammers, victims, and accomplices. For instance, extensive ties to other scammers or participation in shared schemes can uncover hidden threats and vulnerable users. This paper demonstrates a methodology to diagnose Scam propagation across social network platforms, thereby enhancing online safety for security firms and users.