A critical review of feature selection methods for machine learning in IoT security

In the internet of things (IoT) era, the security of connected devices and systems is critical. Machine learning models are commonly used for IoT attack detection, where feature selection (FS) plays an important role. However, FS for IoT security differs from traditional cybersecurity due to the uniqueness of IoT systems. This paper reviews FS methods for effective machine learning-based IoT attack detection. We identify five research questions and systematically review 1,272 studies, analysing 63 that meet inclusion criteria using the preferred reporting items for systematic review and meta-analysis (PRISMA) guidelines. We categorised the studies to address the research questions regarding FS methods, trends, practices, datasets and validation used. We also discussed FS limitations, challenges, and future research directions for IoT security. The review can serve as a reference for researchers and practitioners seeking to incorporate effective FS into machine learning-based IoT attack detection.