An Efficient Real-Time NIDS Using Machine Learning Methods

Recent developments in network technology and related services have caused a significant rise in data traffic. However, there has also been a massive rise in the negative consequences of cyber-attacks. Many new types of network attacks are emerging. As a result, designing a robust Intrusion detection system (IDS) has become essential. This paper presents a framework for designing an efficient IDS to enhance detection accuracy and reduce false positives on real-time data. This research used the CIC-IDS 2017 dataset to train Machine Learning models such as Logistic Regression, K Nearest Neighbor, Gaussian Naive Bayes, and Random Forest. Machine learning models often perform well on benchmark datasets but may encounter challenges when applied to real-time traffic scenarios. So, we created a Real-time dataset and tested it on the trained models. In the evaluation, the Random Forest classifier outperformed all other models and achieved an accuracy of 99.99%.