Enhancing Intrusion Detection System Performance: A Comprehensive Study on Feature Selection and Machine Learning Algorithms

This research addresses the challenges faced by Intrusion Detection Systems (IDS) in processing vast and diverse data streams generated by modern technologies. By integrating feature selection methods with various machine learning algorithms, the study aims to enhance IDS performance while reducing computational costs. Using a multi-class dataset encompassing benign and common DDoS attacks, the research employs the ANOVA feature selection technique to reduce dimensionality and preserve critical features. Empirical evaluations of various machine learning algorithms, including Random Forest, Logistic Regression, Decision Tree, Stochastic Gradient, Adaboost, K-Nearest Neighbour, Extra Trees, and Naïve Bayes, demonstrate their effectiveness in intrusion detection. Comparative analysis is conducted using performance metrics such as accuracy, precision, recall, and F1score.