Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{192471,
        author = {Bestha Umesh Chandra and Kusumala Divakar and Yeranakula Veerendra and Bannela Kuruva Mallesh and P. Sujatha and Dr. P. Veeresh},
        title = {Simple Yet Powerful Machine Learning-Based IoT Intrusion System With Smart Preprocessing and Feature Generation Rivals Deep Learning},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {9},
        pages = {1219-1228},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=192471},
        abstract = {Recent hype on Internet of Things (IoT) has led to the increased susceptibility of cyber-attacks and that, more than ever, has led to the timeliness of the need to have effective intrusion detection systems (IDS). To construct an IoT intrusion detection network, the paper will apply machine learning on the data of UNSW-NB15 dataset to preclassify the attacks into eight categories, including: DoS, Exploits, Fuzzers, Generic, Normal, Reconnaissance, Shellcode and Worms. Decision Tree (DT), Random Forest (RF), as well as XGBoost, LightGBM, AdaBoost, Artificial Neural Networks (ANN) and Convolutional Neural Networks (CNN) are some of the machine learning models which have been trained and evaluated based on the classification accuracy. Optimization of the hyper parameter was carried out using both models and the outcome of the optimization is the random forest and the rate of 98.20 percent and the LightGBM whose rate is 98.13. The correctness of Decision Tree model and the XGBoost were 97.22 and 94.67, respectively. AdaBoost and ANN had a precision of 71.93 and 86.02 respectively. The above models were regarded as the most precise and the highest score of Relative Recall and F1-score with the highest score of most forms of attacks presumed to be the Random Forest and LightGBM models. It has been established that machine learning algorithms such as the random forest and lightgbm can be implemented to predict the network-based attacks on the network of the IoT devices and this can become a solution to the security of the IoT networks.},
        keywords = {Intrusion Detection, Machine learning, UNSW-NB15 Dataset, Attack Classification, DoS, Exploits, Fuzzers, Generic, Normal, Reconnaissance, Shellcode, Worms, Decision Tree, Random Forest, XGBoost, LightGBM, AdaBoost, Artificial neural network.},
        month = {February},
        }