Feature engineering based intelligent intrusion detection system for IoT network security

Internet-of-Things (IoT) helps create smart systems by allowing physical environments to be controlled or monitored by users through extensive sensor networks. Such networks are heavily resource-constrained due to the low processing power and decentralised placement of sensors. Having been implemented in several backbone sectors such as industry, healthcare, etc., they often deal with sensitive information. Thus, there is an imminent need to secure IoT networks using lightweight intrusion detection systems. A novel feature engineering-based intelligent network intrusion detection framework, dubbed FEIIDS (Feature Engineering based Intelligent Intrusion Detection System) entirely based on the Machine Learning concept is proposed. Its working is explained in depth and its time complexity is analyzed. The model is trained and tested over the UNSW-NB15 dataset and found to be competitive with existing works, achieving an accuracy of up to 98.8% and 97.03% for binary and multiclass classification. The proposed work also attains 99% for precision, recall, and F1 score. On the other hand, state-of-the-art work that uses a bidirectional long short-term memory (BiLSTM) model for network intrusion detection is explored over different run-time hardware environments (CPU, GPU, and TPU). Early stopping is implemented to improve model performance and enable better analysis of the results. Discussions on the results reveal that FEIIDS is more stable, transparent, efficient, and lightweight compared to existing state-of-the-art models.