Bayesian-Based Industrial Internet Service Abnormal Detection Algorithm

Abstract

The large and complex nature of the industrial Internet has led to a large amount of attack information in network traffic, and network attacks will cause network traffic anomalies. How to quickly and accurately detect network traffic anomalies and reduce the labor cost of detection model training has become an important issue in the development of network technology. Aiming at this problem, this paper proposes a Bayesian-based industrial Internet service abnormal detection algorithm. Based on the abnormal detection of LightGBM traffic, Bayesian optimization can further improve the efficiency and accuracy of the algorithm's traffic anomaly detection and reduce the manual participation of model training. The experimental results show that the proposed algorithm can improve the automation degree of the algorithm and reduce the labor cost in the model training based on the effective discovery of abnormal traffic.