An Intelligent CNN–LSTM Hybrid Deep Learning-Based Intrusion Detection Framework for Zero-Day Attack Detection in Network Traffic

Abstract

The rapid growth of digital communication technologies and internet-based services has significantly increased the vulnerability of modern network infrastructures to sophisticated cyber threats and zero-day attacks. Traditional Intrusion Detection Systems often struggle to accurately identify unknown and evolving attack patterns due to their dependency on predefined signatures and handcrafted features. To address these limitations, this study presents a novel Deep Learning-based Intrusion Detection System utilizing a hybrid CNN–LSTM architecture for efficient zero-day attack identification in network traffic environments. The proposed framework integrates the powerful spatial feature extraction capability of Convolutional Neural Networks with the sequential learning and temporal dependency analysis strengths of Long Short-Term Memory networks. The system is designed to automatically learn complex traffic patterns, detect anomalous behaviors, and classify malicious activities with high precision and minimal false alarm rates. Extensive preprocessing techniques, including feature normalization, noise reduction, and traffic balancing, are applied to improve data quality and model performance. The proposed hybrid architecture is evaluated using benchmark cybersecurity datasets to analyze detection efficiency, classification accuracy, precision, recall, F1-score, and computational performance. Experimental findings demonstrate that the CNN–LSTM hybrid model significantly outperforms traditional machine learning approaches and standalone deep learning models in identifying unknown and zero-day cyber threats. The results indicate that the proposed framework achieves enhanced detection accuracy, faster response capability, improved generalization, and better adaptability to dynamic network conditions. Furthermore, the model provides effective real-time intrusion monitoring and strengthens proactive cyber defense mechanisms for modern intelligent networks. This research contributes to the advancement of AI-driven cybersecurity solutions by offering a scalable, robust, and efficient deep learning framework capable of improving network security resilience against emerging cyber attacks and sophisticated intrusion attempts in next-generation communication systems.