Skip to content
Open access

Performance Evaluation of Boosting Algorithms on Intrusion Detection System Based on Real-Time CICIoT2023 Dataset

Sep 2026 · International Islamic University Malaysia Engineering Journal · Vol 27, pp. 250-265 · 0 citations · 22 references

Abstract

Currently, the utilization of machine learning methods for identifying intrusions in Internet of Things (IoT) networks demonstrates intriguing prospects. Choosing the appropriate machine learning technique for intrusion detection poses a significant challenge. Choosing the wrong algorithm may reduce threat-detection accuracy, increase the likelihood of network contamination, and compromise network security. This research conducts a comprehensive assessment of various boosting algorithms for detecting intrusions in IoT, using the latest CICIoT2023 dataset. The approach includes evaluating the effectiveness of these algorithms in both binary and multi-class classification scenarios. The evaluation includes a thorough analysis of classification performance metrics, including accuracy, precision, recall, and F1 score. It also assesses computational efficiency, particularly the duration of the training and testing procedures. The results show that the boosting algorithms have potential on the dataset used. Extreme Gradient Boosting yielded the highest accuracy for multi-class classification, whereas Categorical Boosting was the most effective for binary classification. Additionally, Categorical Boosting demonstrated superior test-time performance for multi-class classification, whereas Gradient Boosting excelled in binary classification. ABSTRAK: Pada masa ini, penggunaan kaedah pembelajaran mesin bagi mengenal pasti pencerobohan rangkaian Internet Benda (IoT) menunjukkan potensi menarik. Memilih teknik pembelajaran mesin yang sesuai bagi mengesan pencerobohan menimbulkan cabaran ketara. Memilih algoritma yang salah boleh mengakibatkan ketepatan berkurangan dalam mengesan ancaman, kemungkinan pencemaran rangkaian yang lebih tinggi, dan keselamatan rangkaian terjejas. Kajian ini menjalankan penilaian komprehensif terhadap pelbagai algoritma penggalak yang digunakan dalam mengesan pencerobohan dalam IoT, menggunakan set data CICIoT2023 terkini. Pendekatan ini merangkumi penilaian keberkesanan algoritma ini dalam senario pengelasan binari dan berbilang kelas. Penilaian ini merangkumi analisis menyeluruh metrik prestasi pengelasan, termasuk ketepatan, penarikan balik, dan skor F1. Selain itu, ia melibatkan penilaian kecekapan pengiraan, terutama pada tempoh prosedur latihan dan pengujian. Keputusan menunjukkan bahawa algoritma penggalak mempunyai potensi pada set data yang digunakan. Pendekatan Galakan Kecerunan Ekstrem menghasilkan tahap ketepatan tertinggi bagi pengelasan berbilang kelas, manakala Galakan Kategori ditunjukkan sebagai yang paling berkesan bagi pengelasan binari. Di samping itu, algoritma Galakan Kategori menunjukkan prestasi masa ujian yang unggul bagi pengelasan berbilang kelas, manakala algoritma Galakan Kecerunan adalah terbaik pada pengelasan binari.

Read PDF

Similar papers

Open access Sep 2026

TOWARD OPERATIONALLY RELIABLE IOT NETWORK INTRUSION DETECTION: EXPLAINABLE BOOSTING WITH DIAGNOSTIC ERROR ANALYSIS

These findings demonstrate that integrating explainable AI with boosting algorithms achieves near-optimal detection performance while revealing structural dataset biases and operational blind spots, and the proposed framework supports the development of interpretable, high-performance IoT intrusion detection systems.

Enoch Success Boakai, Sunil Lalchand Bajeja · 0 citations
Open access Sep 2026

Enhanced Intrusion Detection System (IDS) Using Machine Learning

The findings support the adoption of hybrid IDS architectures as a balanced and practical solution that enhances detection capability, adaptability, and reliability in evolving cyber threat landscapes.

Bang-Chen Yu · 0 citations
Review Open access Oct 2026

A Systematic Review of Machine Learning Methods for Network Intrusion Detection: Performance, Datasets, and Cybersecurity Implementation Challenges

The increasing frequency and complexity of cyberattacks have highlighted the need for effective Network Intrusion Detection Systems (NIDS). Machine learning (ML) has emerged as a promising approach for identifying malicious network traffic and detecting anomalous activities. However, differences in algorithms, datasets...

Ridwan Ridwan, Fivi Aprilianingsih · 0 citations
Open access Oct 2026

Explainable Machine Learning for Real-Time Cyber Threat Detection: An Intelligent Framework for Adaptive Information Security

The cyber threats in IoT environments are constantly changing and need accurate, interpretable, efficient and adaptive intrusion detection. This research suggests an explainable machine-learning framework with the CICIoT2023 dataset. Logistic Regression, Decision Tree, Random Forest, XGBoost, and Explainable Boosting M...

Sunilgar Gusai, Shrikant Telang, R. Balakrishnan et al. · 0 citations
Conference Open access Oct 2026

Explainable Cybersecurity Intrusion Detection Framework using Machine Learning and Ensemble Models

Digital communication systems are rapidly growing, and this rapid growth has also brought a lot of risk of attacks and unauthorized network intrusion, making the need for intelligent intrusion detection extremely high. Efficiently detecting complex and evolving attack patterns is often a challenge with traditional appr...

S. Vairachilai, R. Vijayalakshmi, T. Yamuna et al. · 0 citations
Oct 2026

Cyber Threats and Network Intrusion Detection: A Comparative Analysis of Soft and Hard Voting Ensemble Methods

A thorough comparison of soft and hard voting ensemble techniques for binary network traffic classification is provided in this article, and the suggested approach combines five distinct supervised machine learning (ML) techniques to create a powerful classifier system.

M. Alshaikh, Farid Binbeshr, Jumana Alamoudi et al. · 0 citations

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.