Modelling graph dynamics in fraud detection with "Attention"

TL;DR

This paper introduces DyHGN, a novel dynamic heterogeneous graph neural network that effectively captures temporal and diverse information for fraud detection on online retail platforms, enhancing explainability and model understanding.

Contribution

It proposes DyHGN and variants for modeling dynamic heterogeneous graphs, incorporating diachronic embeddings and graph transformers for fraud detection.

Findings

Modeling graph dynamics with attention improves fraud detection accuracy.

Heterogeneous and temporal data require attention-based approaches for better performance.

Explainability techniques help understand model behaviors in complex graph scenarios.

Abstract

At online retail platforms, detecting fraudulent accounts and transactions is crucial to improve customer experience, minimize loss, and avoid unauthorized transactions. Despite the variety of different models for deep learning on graphs, few approaches have been proposed for dealing with graphs that are both heterogeneous and dynamic. In this paper, we propose DyHGN (Dynamic Heterogeneous Graph Neural Network) and its variants to capture both temporal and heterogeneous information. We first construct dynamic heterogeneous graphs from registration and transaction data from eBay. Then, we build models with diachronic entity embedding and heterogeneous graph transformer. We also use model explainability techniques to understand the behaviors of DyHGN-* models. Our findings reveal that modelling graph dynamics with heterogeneous inputs need to be conducted with "attention" depending on the data structure, distribution, and computation cost.