Edge-Based DDoS Attack Detection in AIoT Using a Hybrid CNN and Logistic Regression Approach

This study presented an edge-based anomaly detection model for AIoT environments using hybrid CNN and Logistic Regression. The model has been evaluated against KDD dataset, and it can be found that it can differentiate DDoSlDoS attack traffic from normal traffic with 94 % accuracy. In addition, it can be noted that the developed model demonstrates the level of performance, which can be considered as satisfactory, since it has the precision of 0.94, the recall of 0.94, and f1-score of 0.94. Thus, it can be stated that the developed edge-based anomaly detection model can be successfully implemented for ensuring the real-time network security and detecting, and preventing the anomalies in the various smart home applications based on its high potential.