What question did this study set out to answer?

The aim is to develop a machine learning strategy for accurately detecting anomalies in IoT devices.

March 10, 2026

Framework for IoT Devices Anomaly Mitigation Strategy Using Machine Learning Technique

ZFZaheen FatimaHamdard University RHRashid HussainHamdard University ADAzhar DilshadHamdard University

Key Points

The aim is to develop a machine learning strategy for accurately detecting anomalies in IoT devices.
Utilized various machine learning algorithms including Logistic Regression, Decision Tree, Random Forest, K-Nearest Neighbors, Multilayer Perceptron, and XGBoost.
Conducted comprehensive analysis on datasets including iot_intrusion and iotid20.
Focused on enhancing the classification performance for distinguishing authorized and unauthorized users.
Achieved promising classification accuracy across different machine learning algorithms.
Demonstrated effectiveness of the proposed framework in mitigating anomalies in IoT systems.
Highlighted operational suitability for handling large datasets.

Abstract

ABSTRACT The Internet of Things (IoT) ecosphere encompasses web‐based devices that have Sensors, actuators, processors, controllers, and data connectivity equipment. These devices collect and exchange data wirelessly through web‐based technologies. IoT devices expedite the man‐to‐machine and machine‐to‐machine interactions. There is always a possibility of an attack on IoT‐enabled network devices due to limited security mechanisms. Devices in the IoT Eco‐system are connected/disconnected frequently, therefore it opens the door for intruders to steal sensitive information and deceptively affect the devices. It seems essential as time passes to present a good detection mechanism by classifying the authorized/unauthorized user. This research aims to implement the machine learning algorithm with a comprehensive analysis that augments the classification and accurately mitigates the anomaly in the IoT system. The achieved promising results show the effectiveness of the machine learning algorithms Logistic Regression (LR), Decision Tree (DT), Random Forest (RF), K‐Nearest Neighbors (KNN), Multilayer Perceptron (MLP), and Extreme Gradient Boosting (XGBoost) with a classification mechanism to enhance the performance by training the model with state of art algorithms which have good predicting capability that improves accuracy in classification mechanisms and is very operational for large datasets. The datasets used for this research are iotᵢntrusion and iotid20.

AIに質問

Bookmark

View Full Paper