The paper introduces a novel method for credit card fraud detection, the Causal Temporal Graph Neural Network (CaT-GNN). This method leverages causal invariant learning to reveal inherent correlations within transaction data. The problem is decomposed into discovery and intervention phases. CaT-GNN identifies causal nodes within the transaction graph and applies a causal mixup strategy to enhance the model’s robustness and interpretability. The method consists of two key components: Causal-Inspector and Causal-Intervener. Evaluated on three datasets, CaT-GNN demonstrates superior performance over existing state-of-the-art methods.
Publication date: 23 Feb 2024
Project Page: Not provided
Paper: https://arxiv.org/pdf/2402.14708
