An Enhanced Machine Learning-Based Intrusion Detection System for IoT Networks Using Feature Engineering and Data Preprocessing
International Journal of Advanced Research in Science Communication and Technology · 22.11.2026 · 381
DOI: 10.48175/ijarsct-40945 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Advanced Research in Science Communication and Technology, online erschienen am 22.11.2026. Verfasst von Vaishnavi Nandu Shelke und 3 weiteren Autor:innen. Die Arbeit nennt 8 Literaturangaben.
Abstract (Originalsprache)
The rapid expansion of Internet of Things (IoT) devices has created new opportunities for smart and connected applications, but it has also increased the number of potential security vulnerabilities. IoT networks generate large volumes of heterogeneous traffic, making the detection of malicious activities a challenging task. Intrusion Detection Systems (IDS) based on machine learning can identify abnormal network behaviour and provide an effective security mechanism for IoT environments. However, the performance of a machine learning model strongly depends on the quality of the input data, preprocessing techniques, feature representation, and class distribution. This research focuses on developing a machine learning-based IoT intrusion detection system with particular emphasis on data preprocessing and feature generation. The proposed framework includes data cleaning, handling of missing values, encoding of categorical attributes, feature reduction, and feature generation before applying a machine learning classifier. The ToN-IoT dataset is considered for evaluating the proposed approach because it contains normal and malicious IoT network traffic representing different attack categories. Random Forest is considered as the primary classification algorithm because of its ability to handle high-dimensional data, nonlinear relationships, and multiple classes while maintaining relatively low computational complexity. Appropriate techniques for handling class imbalance are also incorporated to improve the detection of minority attack categories. The main objective of this study is to investigate whether a carefully designed preprocessing and feature engineering pipeline can improve the performance of a classical machine learning model for IoT intrusion detection. The study also considers important evaluation measures such as accuracy, precision, recall, F1-score, and ROC-AUC to provide a comprehensive assessment of detection performance. The proposed approach aims to provide an efficient, interpretable, and practical alternative to highly complex deep learning-based intrusion detection systems, particularly for resource-constrained IoT environments.
Zitieren
APA
Shelke, V. N., Shimpi, P. J. K., Satre, P. S. D., & Jadhav, D. H. B. (2026). An Enhanced Machine Learning-Based Intrusion Detection System for IoT Networks Using Feature Engineering and Data Preprocessing. International Journal of Advanced Research in Science Communication and Technology, 381. https://doi.org/10.48175/ijarsct-40945
Vancouver
Shelke VN, Shimpi PJK, Satre PSD, Jadhav DHB. An Enhanced Machine Learning-Based Intrusion Detection System for IoT Networks Using Feature Engineering and Data Preprocessing. International Journal of Advanced Research in Science Communication and Technology. 2026:381. doi:10.48175/ijarsct-40945
BibTeX
@article{Shelke2026,
author = {Shelke, Vaishnavi Nandu and Shimpi, Prof. J. K. and Satre, Prof. S. D. and Jadhav, Dr. H. B.},
title = {An Enhanced Machine Learning-Based Intrusion Detection System for IoT Networks Using Feature Engineering and Data Preprocessing},
journal = {International Journal of Advanced Research in Science Communication and Technology},
year = {2026},
pages = {381},
doi = {10.48175/ijarsct-40945}
}
EUVIMED