Chinese Fraudulent Text Message Detection Based on Graph Neural Networks
Peiwen Gao, Liang Zhang · 2024
With the widespread adoption of mobile phones, the frequency of criminals exploiting text messages for scams is on the rise. However, due to the sparse and deceptive nature of Chinese text messages, existing detection methods struggle to effectively combat this issue. To address this challenge, we introduce a novel network model leveraging graph neural networks. Initially, we utilize BERT as a word embedding tool to derive word vectors from the text. Subsequently, we construct a word-document heterogeneous graph based on these word vectors. Then, employing Graph Attention Networks (GAT), we learn features between the nodes of the graph. Finally, we integrate the predictions from BERT and GAT to form the final prediction result. Experimental results on two publicly available datasets demonstrate that our proposed model outperforms other state-of-the-art models in performance.