Hybrid ML and DL Framework for Advanced Network Intrusion Detection and Prevention
S. Palani, A. Muthukumaravel · 2025
Such requirements are brought on by the dynamics of cyber threats. In the current study, an enhanced version of Network Intrusion Detection and Prevention System is proposed using hybrid architecture that combines Machine Learning and Deep Learning methods for intrusion detection purposes. To improve input data as well as decrease computational cost and complexity, feature selection based on ML is considered for the proposed approach. Network abnormality identification and classification is well done by Deep Learning architecture, and for its accuracy, this paper used CNN combined with Long Short-Term Memory. That is to say, it applies ML in collaboration with DL which increases detection but significantly reduces the false positives in the output of the detection result by maintaining balance between efficiency and accuracy. The framework, which has an accuracy of over 98% and powerful real-time capabilities, proves to be the best in experimental assessments conducted on benchmark datasets. Secure network systems future advancements may be based on this hybrid approach, offering an adaptable and scalable answer to the problem of sophisticated network invasions.