This paper compares feature selection and feature extraction techniques for optimizing IoT intrusion detection systems (NIDS) using machine learning. The study focuses on the Network TON-IoT dataset, which includes both binary and multiclass classification tasks. Key performance metrics such as accuracy, F1-score, and runtime are evaluated to assess the effectiveness of these methods. Feature selection reduces the dimensionality of the data by selecting a subset of informative features, while feature extraction transforms the original features into a lower-dimensional space. The results show that feature extraction generally performs better in terms of detection accuracy, especially when the number of features is small, and it is less sensitive to changes in the number of features. However, feature selection achieves faster model training and inference times. The study provides guidelines for selecting appropriate intrusion detection methods based on the specific requirements and constraints of IoT systems. The findings highlight the importance of balancing detection accuracy and computational complexity in IoT NIDS.This paper compares feature selection and feature extraction techniques for optimizing IoT intrusion detection systems (NIDS) using machine learning. The study focuses on the Network TON-IoT dataset, which includes both binary and multiclass classification tasks. Key performance metrics such as accuracy, F1-score, and runtime are evaluated to assess the effectiveness of these methods. Feature selection reduces the dimensionality of the data by selecting a subset of informative features, while feature extraction transforms the original features into a lower-dimensional space. The results show that feature extraction generally performs better in terms of detection accuracy, especially when the number of features is small, and it is less sensitive to changes in the number of features. However, feature selection achieves faster model training and inference times. The study provides guidelines for selecting appropriate intrusion detection methods based on the specific requirements and constraints of IoT systems. The findings highlight the importance of balancing detection accuracy and computational complexity in IoT NIDS.