Comparative analysis of deep learning and traditional methods for IoT botnet detection using a multi-model framework across diverse datasets

Abstract The proliferation of Internet of Things (IoT) devices has created unprecedented cybersecurity vulnerabilities, with botnets emerging as a critical threat to network infrastructure. This study focuses on traditional machine learning and deep learning approaches, proposes a novel ensemble fra...

Full description

Saved in:
Bibliographic Details
Main Authors: Saeed Ullah, Junsheng Wu, Zhijun Lin, Mian Muhammad Kamal, Hala Mostafa, Muhammad Sheraz, Teong Chee Chuah
Format: Article
Language:English
Published: Nature Portfolio 2025-08-01
Series:Scientific Reports
Subjects:
Online Access:https://doi.org/10.1038/s41598-025-16553-w
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849226254898692096
author Saeed Ullah
Junsheng Wu
Zhijun Lin
Mian Muhammad Kamal
Hala Mostafa
Muhammad Sheraz
Teong Chee Chuah
author_facet Saeed Ullah
Junsheng Wu
Zhijun Lin
Mian Muhammad Kamal
Hala Mostafa
Muhammad Sheraz
Teong Chee Chuah
author_sort Saeed Ullah
collection DOAJ
description Abstract The proliferation of Internet of Things (IoT) devices has created unprecedented cybersecurity vulnerabilities, with botnets emerging as a critical threat to network infrastructure. This study focuses on traditional machine learning and deep learning approaches, proposes a novel ensemble framework to address these issues, integrating Convolutional Neural Network (CNN), Bidirectional Long Short-Term Memory (BiLSTM), Random Forest (RF), and Logistic Regression (LR) via a weighted soft-voting mechanism. Our approach introduces a Quantile Uniform transformation to reduce feature skewness, a multi-layered feature selection method to enhance discriminative power, an individual performance of deep learning–traditional machine learning and a hybrid models (ensemble models) for robust detection. Evaluated on BOT-IOT, CICIOT2023, and IOT23 datasets, the framework achieves 100% accuracy on BOT-IOT, 99.2% on CICIOT2023, and 91.5% on IOT23, outperforming state-of-the-art models by up to 6.2%. These contributions advance IoT security by enabling scalable, high-performance detection adaptable to diverse network scenarios, with practical optimizations for real-world deployment.
format Article
id doaj-art-0c630eba89164a24b0bd16219f09d6ff
institution Kabale University
issn 2045-2322
language English
publishDate 2025-08-01
publisher Nature Portfolio
record_format Article
series Scientific Reports
spelling doaj-art-0c630eba89164a24b0bd16219f09d6ff2025-08-24T11:27:57ZengNature PortfolioScientific Reports2045-23222025-08-0115113110.1038/s41598-025-16553-wComparative analysis of deep learning and traditional methods for IoT botnet detection using a multi-model framework across diverse datasetsSaeed Ullah0Junsheng Wu1Zhijun Lin2Mian Muhammad Kamal3Hala Mostafa4Muhammad Sheraz5Teong Chee Chuah6School of Software, Northwestern Polytechnical UniversitySchool of Software, Northwestern Polytechnical UniversitySchool of Computer Science, Northwestern Polytechnical UniversitySchool of Electronics and Communication Engineering, Quanzhou University of Information EngineeringDepartment of Information Technology, College of Computer and Information Sciences, Princess Nourah bint Abdulrahman UniversityCentre for Smart Systems and Automation, CoE for Robotics and Sensing Technologies, Faculty of Artificial Intelligence and Engineering, Multimedia UniversityCentre for Smart Systems and Automation, CoE for Robotics and Sensing Technologies, Faculty of Artificial Intelligence and Engineering, Multimedia UniversityAbstract The proliferation of Internet of Things (IoT) devices has created unprecedented cybersecurity vulnerabilities, with botnets emerging as a critical threat to network infrastructure. This study focuses on traditional machine learning and deep learning approaches, proposes a novel ensemble framework to address these issues, integrating Convolutional Neural Network (CNN), Bidirectional Long Short-Term Memory (BiLSTM), Random Forest (RF), and Logistic Regression (LR) via a weighted soft-voting mechanism. Our approach introduces a Quantile Uniform transformation to reduce feature skewness, a multi-layered feature selection method to enhance discriminative power, an individual performance of deep learning–traditional machine learning and a hybrid models (ensemble models) for robust detection. Evaluated on BOT-IOT, CICIOT2023, and IOT23 datasets, the framework achieves 100% accuracy on BOT-IOT, 99.2% on CICIOT2023, and 91.5% on IOT23, outperforming state-of-the-art models by up to 6.2%. These contributions advance IoT security by enabling scalable, high-performance detection adaptable to diverse network scenarios, with practical optimizations for real-world deployment.https://doi.org/10.1038/s41598-025-16553-wIoTCyber-attacksBotnetDeep learningTraditional machine learningSMOTE
spellingShingle Saeed Ullah
Junsheng Wu
Zhijun Lin
Mian Muhammad Kamal
Hala Mostafa
Muhammad Sheraz
Teong Chee Chuah
Comparative analysis of deep learning and traditional methods for IoT botnet detection using a multi-model framework across diverse datasets
Scientific Reports
IoT
Cyber-attacks
Botnet
Deep learning
Traditional machine learning
SMOTE
title Comparative analysis of deep learning and traditional methods for IoT botnet detection using a multi-model framework across diverse datasets
title_full Comparative analysis of deep learning and traditional methods for IoT botnet detection using a multi-model framework across diverse datasets
title_fullStr Comparative analysis of deep learning and traditional methods for IoT botnet detection using a multi-model framework across diverse datasets
title_full_unstemmed Comparative analysis of deep learning and traditional methods for IoT botnet detection using a multi-model framework across diverse datasets
title_short Comparative analysis of deep learning and traditional methods for IoT botnet detection using a multi-model framework across diverse datasets
title_sort comparative analysis of deep learning and traditional methods for iot botnet detection using a multi model framework across diverse datasets
topic IoT
Cyber-attacks
Botnet
Deep learning
Traditional machine learning
SMOTE
url https://doi.org/10.1038/s41598-025-16553-w
work_keys_str_mv AT saeedullah comparativeanalysisofdeeplearningandtraditionalmethodsforiotbotnetdetectionusingamultimodelframeworkacrossdiversedatasets
AT junshengwu comparativeanalysisofdeeplearningandtraditionalmethodsforiotbotnetdetectionusingamultimodelframeworkacrossdiversedatasets
AT zhijunlin comparativeanalysisofdeeplearningandtraditionalmethodsforiotbotnetdetectionusingamultimodelframeworkacrossdiversedatasets
AT mianmuhammadkamal comparativeanalysisofdeeplearningandtraditionalmethodsforiotbotnetdetectionusingamultimodelframeworkacrossdiversedatasets
AT halamostafa comparativeanalysisofdeeplearningandtraditionalmethodsforiotbotnetdetectionusingamultimodelframeworkacrossdiversedatasets
AT muhammadsheraz comparativeanalysisofdeeplearningandtraditionalmethodsforiotbotnetdetectionusingamultimodelframeworkacrossdiversedatasets
AT teongcheechuah comparativeanalysisofdeeplearningandtraditionalmethodsforiotbotnetdetectionusingamultimodelframeworkacrossdiversedatasets