Optimizing Network Security: A CRNN Architecture for Firewall Anomaly Detection
Asfiya Shireen Shaikh Mukhtar, R. N. Jugele · 2025
Continuous development in security technologies is essential to address the increasingly complex nature of assaults in the rapidly changing area of cybersecurity. Despite they've proven effective in times gone by, conventional strategies have become unable to keep up with the ever-changing nature of intrusions. For the purpose to apply convolutional neural networks (CNNs) to identify firewall anomalies. This study offers an original design that systematically encompasses three deep learning (DL) models and seven machine learning (ML) models. The suggested framework combines CNN_LSTM, Feedforward Neural Network, Neural Network, Neighbours Classifier, Gaussian NB, Linear SVC, and Random Forest Classifier over depending on rule-driven approaches. employing machine learning to offer the firewall intelligence and responsiveness, the system gets over the restrictions associated with conventional firewalls and enabling it to identify and react to constantly changing malware autonomously. The CNN-LSTM composite model's spatial-temporal sensitivity is further demonstrated by its incorporation of deep neural network models, which enhance the architectural capacity to recognize complex structures. This interdisciplinary research seeks to redefine the efficient functioning of security measures versus the evolving cyber-attack environment by improving safety measures and contributing to the broader discussions on employing machine learning and deep learning in real-world situations.