Performance Analysis of Intrusion Detection System in the IoT Environment Using Feature Selection Technique

• DOI: 10.32604/iasc.2023.036856
• Publication Date: 2023-03-15T03:44:52Z
• Authors: Moody Alhanaya; Khalil Hamdi Atey...
• Keywords: Feature (linguistics); Information gain ratio

The increasing number of security holes in the Internet Things (IoT) networks creates a question about reliability existing network intrusion detection systems. This problem has led to developing research area focused on improving network-based system (NIDS) technologies. According analysis different businesses, most researchers focus classification results NIDS datasets by combining machine learning and feature reduction techniques. However, these techniques are not suitable for every type network. In light this, whether optimal algorithm can be generalized across various IoT remains. paper aims analyze methods used this they other datasets. Six ML models were study, namely, logistic regression (LR), decision trees (DT), Naive Bayes (NB), random forest (RF), K-nearest neighbors (KNN), linear SVM. primary algorithms Principal Component (PCA) Gini Impurity-Based Weighted Forest (GIWRF) evaluated against three global ToN-IoT datasets, UNSW-NB15, Bot-IoT dimensions each dataset was studied applying PCA algorithm. It is stated that selection affects performance FE used. Increasing efficiency requires comprehensive standard set improve quality over time.