ANALYSE THE EFFECTS OF NORMALIZATION ON NSL-KDD INTRUSION DETECTION SYSTEM

Protecting unauthorized access is critical in the realm of information and network security for everyone's well-being. The Intrusion Detection System (IDS) is a type of intrusion detection system that plays one of the most significant roles here. It's a classifier that determines if the data is normal or malicious. In this paper, the author used several Machine Learning Techniques for Intrusion Detection to investigate the effects of normalization on the NSL-KDD dataset. In this experiment, WEKA is an open-source data mining tool that we utilized. The Decision Tree, Naïve Bayes, Random Forest, and One R algorithms are applied in the context of with and without normalization. Different normalization techniques are used, such as, Min-Max, Z-Score, Log Scaling, and Mean Centered Scaling. The obtained result shows that the Random Forest has a greater accuracy rate than others. In both cases, i.e., with and without normalization.