Enhancing Intrusion Detection with CNN-RF Hybrid Model: A High-Performance Approach Using NSL-KDD Dataset

The growing sophistication of attacks is accompanied by the growing significance of having developed sophisticated intrusion detection systems (IDS) that can accurately determine the threats. This paper thus suggests improving the identification of network intrusions by combining RF with CNN. The model depends on the hierarchical feature extraction by CNN for the classification and ensemble learning of RF for the selection of features. In terms of the experimental results obtained, the proposed model beats traditional IDS strategies when tested on the NSL-KDD dataset with a recall of 0.96, an F1-score of 0.965, an accuracy of 98.73%, and a precision of 0.97. The use of this more complex data preprocessing method provides a guarantee that it would be able to detect unique types of attacks and is more resilient to imbalanced datasets. The technology is adaptable as well as scalable, showcasing the potential to be developed and applied for real-time intrusion detection.