Unified representation and scoring framework for anomaly detection in attributed networks with emphasis on structural consistency and attribute integrity

Abstract Anomaly detection in attributed networks has attracted considerable interest owing to its applications in security, fraud detection, and system monitoring. These networks pose unique challenges as anomalies may arise from either structural irregularities or inconsistencies in node attribute...

Full description