A Comprehensive ML Framework for Intrusion Detection

In order to identify and stop network intrusions in their companies and to stop network breaches, security analysts and administrators must overcome numerous obstacles. Timely detection of the intrusion is essential. Principal component analysis (PCA) is used for dimensionality reduction, and ant colony optimization (ACO) is used for feature selection. Classification is done using the Extreme Gradient Boost Algorithm (XG-Boost) using the NSL-KDD dataset. To find the ideal hyperparameters and network configurations for machine learning models, numerous trial and error experiments were conducted. Performance metrics calculated are 98.2% Accuracy, 99.2% Precision, 97.1% Recall and 98.1% F1Score.