Scalable Network Intrusion Detection using Hybrid RNN-MLP Architectures

There is an elevated rise of threats against network infrastructure in this digital age through cyber threats. This cause makes the risk of malicious network intrusions elevate and higher for organizations and individuals. This paper discusses Network Intrusion Detection (NID) utilizing deep learning to differentiate benign and malicious network traffic. A number of performances of algorithms were evaluated MLP, Random Forest, and LSTM-based recurrent neural networks. The results of evaluation show that the proposed solution is appropriate; therefore, the highest accuracy was achieved by the MLP model with 98.49%, which outperformed the others in the related metrics of accuracy, precision, recall, and F1-score. Therefore, it aims to provide organizations with a robust network security solution that will minimize the impact of cyber intrusions against their network infrastructures while ensuring integrity and safety.