A Universal Adaptive Algorithm for Graph Anomaly Detection

Yuqi Li, Guosheng Zang, Chunyao Song, Xiaojie Yuan

Abstract: Highlights•Introduced k<math><mi is="true">k</mi></math>FN-graph for camouflage anomaly detection.•Proposed adaptive graph filter to enhance signals.•Developed U-A2<math><msup is="true"><mrow is="true"><mi is="true">A</mi></mrow><mrow is="true"><mn is="true">2</mn></mrow></msup></math>GAD, combining spectral, kNN, kFN with attention.•Achieved up to 6% AUC-PR improvement over S0TA on real datasets.