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.
Yi-Fan Lu, Qi-Xiao Lin, Jian Mao et al.
· 0 citations