DESIGN AND IMPLEMENTATION OF A MACHINE LEARNING-BASED NETWORK INTRUSION DETECTION SYSTEM WITH COMPARATIVE ANALYSIS OF CLASSIFICATION ALGORITHMS

Intrusion Detection Systems (IDS) are vital for network security of computers as they detect attacks and unauthorized access to networks.Signature based IDS are difficult for detecting changing attacks, so Machine Learning based is a good choice of technology that can be used as well.This paper describes the design and implementation of a machine learning based Intrusion Detection System using the KDD Cup 1999 (KDD'99) dataset.Data were cleaned up by way of selecting features; encoding categories as well as normalizing them for better results with less computation time taken out from processing them.We chose 13 major characteristics of a networking system to train our models on.Four machine learning algorithms -decision Tree, Random Forest, Gradient Boosting, and Artificial Neural Network (ANN)were developed and evaluated using accuracy, precision, recall, F1-score, and confusion matrix analysis.Experimental results demonstrate that the Random Forest classifier achieved the highest overall performance with an accuracy of approximately 99%, while also maintaining high detection rates across multiple attack categories, including Denial-of-Service (DoS), Probe, Remote-to-Local (R2L), and User-to-Root (U2R).For demonstration of its utility, we have used an instance of our trained random forest classifier as part of a flask-based web app which enables classification on networks' connections & attacks logs at runtime.This approach provides a good, convenient way to detect intrusions as well; it also points out that ml can be used better at improving cybersecurity on networks.