Comparative Analysis of Machine Learning Models for Intrusion Detection in Internet of Things Networks - Security Performance Evaluation

Aug 5, 2026

Upload by Eden

This research investigates the performance of machine learning models for an intrusion detection system (IDS) in the evolving security landscape of Internet of Things (IoT) networks. Using the RT-IoT2022 dataset, Gradient Boosting and Random Forest achieve perfect results (accuracy, precision, recall, and F1 of 1.00), while Logistic Regression shows consistent scores of 0.96 across metrics. Multi-Layer Perceptron reaches 0.99, indicating strong capability for complex nonlinear patterns, and the study highlights overfitting concerns and the need for continuous evaluation with real-world data.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages7

Reading Time18 min

AccessFree to download

Abstract

1. Introduction

IoT growth and security vulnerabilities

Limitations of traditional IDS

Machine learning approaches for adaptive IDS

2. Research objective and dataset

Comparative Analysis of Machine Learning Models for Intrusion Detection in Internet of Things Networks Using the RT-IoT2022 Dataset Gregorius Airlangga Information System Study Program, Atma Jaya Catholic University of Indonesia, Indonesia E-Mail: [gregorius.airlangga@atmajaya.ac.id](gregorius.airlangga@atmajaya.ac.id) Received Jan 3rd 2024; Revised Feb 25th 2024; Accepted Mar 17th 2024 Corresponding Author: Gregorius Airlangga Abstract This research investigates the performance of various machine learning models in developing an Intrusion Detection System (IDS) for the complex and evolving security landscape of Internet of Things (IoT) networks. Employing the RTIoT2022 dataset, which captures a diverse array of IoT devices and attack methodologies, we meticulously evaluated four prominent models: Gradient Boosting, Random Forest, Logistic Regression, and Multi-Layer Perceptron (MLP). Our results indicate that both Gradient Boosting and Random Forest achieved perfect scores with an accuracy, precision, recall, and F1 score of 1.00, suggesting their superior ability to classify and predict security incidents within the dataset. However, such perfection raises concerns about overfitting, which necessitates further investigation. Logistic Regression demonstrated commendable consistency with scores of 0.96 across all metrics, proposing a balance between model complexity and performance. The MLP model closely followed, with an accuracy, precision, recall, and F1 score of 0.99, highlighting its potential in capturing complex, nonlinear data relationships. These findings underscore the critical role of machine learning in fortifying IoT networks against cyber threats and the need for continuous model evaluation against real-world data. The study provides a pathway for future research to refine these IDS models for operational efficiency and sustainability in the dynamic IoT security domain. Through this work, we aim to contribute to the advancement of secure and resilient IoT infrastructures. Keyword: Cyber Threat Detection, Internet of Things (IoT) Security, Intrusion Detection Systems (IDS), Machine Learning Models, Network Traffic Classification 1. INTRODUCTION In the digital age, the proliferation of Internet of Things (IoT) devices has transformed everyday life, embedding intelligence into our homes, workplaces, and urban spaces [1]–[3] . This transformation, while bringing unparalleled convenience and efficiency, also introduces a plethora of security vulnerabilities [4]–[6] . IoT devices, often designed with limited attention to security, become prime targets for cyberattacks, threatening user privacy, data integrity, and overall network security [7]–[9] . The complexity and diversity of IoT ecosystems further exacerbate these challenges, necessitating the development of sophisticated Intrusion Detection Systems (IDS) that can effectively safeguard these interconnected environments [10]–[12] . The literature on cybersecurity in IoT networks underscores the escalating arms race between attackers and defenders [13]–[15] . Studies such as those by [16] have documented the evolving landscape of IoT threats, including DDoS attacks, malware infiltration, and sophisticated phishing campaigns. These works highlight the limitations of traditional IDS solutions, which often struggle with the dynamic, heterogeneous nature of IoT networks and the novel attack vectors introduced by emerging technologies [17] . Recent research has increasingly focused on leveraging machine learning (ML) and deep learning (DL) techniques to build adaptive IDS that can recognize and mitigate both known and unknown threats [18]–[20] . For instance, [21] demonstrated the potential of distributed deep learning models in detecting distributed network attacks, while [22] provided a comprehensive review of deep learning-based IDS for IoT, identifying key challenges such as model scalability, data imbalance, and real-time detection capabilities.

FAQs

Which machine learning models are evaluated for intrusion detection in IoT networks?

The study evaluates Gradient Boosting, Random Forest, Logistic Regression, and Multi-Layer Perceptron (MLP) using the RT-IoT2022 dataset.

How do Gradient Boosting and Random Forest perform on the dataset?

Both models achieve perfect scores with accuracy, precision, recall, and F1 of 1.00, indicating top classification and prediction performance on the dataset.

What concern is raised about the perfect performance results?

Perfect results may indicate overfitting, so the paper calls for further investigation beyond the initial evaluation.

Aug 5, 2026

Upload by Eden