Hybrid Intrusion Detection System (IDS) Using Machine Learning and Deep Learning

This work offers a Hybrid Intrusion Detection System (HIDS) that combines traditional machine learning and deep learning methods for efficient and scalable network attack identification. The system makes use of Principal Component Analysis (PCA) for reducing dimensionality and then utilizes a hybrid CNN-LSTM architecture for feature learning as well as classification. An ensemble method is also utilized to combine Random Forest with the CNN-LSTM to add robustness as well as generalization. The CICIDS2018 dataset, comprising modern real-world network traffic situations, is employed for testing. Our system detects with an accuracy of 99.1% on the test set, far better than conventional classifiers. This paper proves the efficacy of integrating statistical feature engineering with deep sequential models and ensemble techniques to counter cybersecurity attacks in real-time settings.