Anomaly Detection Based on Behavior Feature Correlation for IoT Systems
The approach models inter-device state correlations using a heterogeneous graph structure and partitions behavior patterns through iterative community detection and automated semantic annotation, and represents normal behavior by embedding and clustering of state sequences.
Yifan Lu, Qixiao Lin, Jian Mao et al.
· 0 citations