Open access
2026
Beyond Simple Aggregation: Decoupling Frequency-Aware Representation Learning and Gradient Boosting for Graph Anomaly Detection
Results show that supervised GAD can be cast as a coordinated process of structural knowledge encoding, adaptive frequency enhancement, and discriminative knowledge learning, and improves AUPRC over the strong baseline built from XGBoost and fixed neighborhood aggregation.
Yangfan Hu, Tingli Yan
· IEEE Access · 0 citations