A traffic anomaly detection approach based on unsupervised learning for industrial cyber-physical system

Tao Yang, Zhenze Jiang, Peiyu Liu, Qiang Yang, Wenhai Wang

Abstract: Highlights•An unsupervised word segmentation model for payloads is developed to accurately segment payloads into words with preserving semantic correlations.•An unsupervised classification model based on an autoencoder is proposed to effectively analyze complex relationships in payloads.•The numerical results show that the proposed detection approach achieves an overall improvement of 20.60% for F1, compared with the state-of-the-art detection approach.