Claim Missing Document
Check
Articles

Found 3 Documents
Search

Attacks Detection in Internet of Things Using Machine Learning Techniques: A Review Amer Dawood Saleem; Amer Abdulmajeed Abdulrahman
Journal of Applied Engineering and Technological Science (JAETS) Vol. 6 No. 1 (2024): Journal of Applied Engineering and Technological Science (JAETS)
Publisher : Yayasan Riset dan Pengembangan Intelektual (YRPI)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.37385/jaets.v6i1.4878

Abstract

The proliferation of IoT devices across sectors such as home automation, business, healthcare, and transportation has led to the generation of vast amounts of sensitive data. This widespread adoption has introduced significant security challenges and vulnerabilities. This study aims to analyze and evaluate machine learning (ML) and deep learning (DL) models for detecting malicious activities in IoT networks, with a focus on improving cybersecurity measures. We conducted a comprehensive review of various ML and DL models, including Random Forest, Decision Tree, HTA-GAN, Hybrid CNN-LSTM, and SVM. The study also includes an evaluation of the datasets used for identifying harmful data, ensuring effective detection of large-scale attacks in IoT ecosystems. Our findings indicate that these models enhance IoT security by deploying efficient intrusion detection systems (IDS) using reliable, large-scale datasets. The study highlights the performance of these models in balancing security and resource management, given the constraints of IoT devices.ML and DL approaches offer significant security benefits for IoT networks, despite the challenges associated with their implementation. The study underscores the importance of future research to address these challenges and further improve IoT security. The results provide valuable insights into the application of ML/DL models in IoT security, contributing to both theoretical knowledge and practical solutions for enhancing cybersecurity in IoT ecosystems.
Attacks Detection in Internet of Things Using Machine Learning Techniques: A Review Amer Dawood Saleem; Amer Abdulmajeed Abdulrahman
Journal of Applied Engineering and Technological Science (JAETS) Vol. 6 No. 1 (2024): Journal of Applied Engineering and Technological Science (JAETS)
Publisher : Yayasan Riset dan Pengembangan Intelektual (YRPI)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.37385/jaets.v6i1.4878

Abstract

The proliferation of IoT devices across sectors such as home automation, business, healthcare, and transportation has led to the generation of vast amounts of sensitive data. This widespread adoption has introduced significant security challenges and vulnerabilities. This study aims to analyze and evaluate machine learning (ML) and deep learning (DL) models for detecting malicious activities in IoT networks, with a focus on improving cybersecurity measures. We conducted a comprehensive review of various ML and DL models, including Random Forest, Decision Tree, HTA-GAN, Hybrid CNN-LSTM, and SVM. The study also includes an evaluation of the datasets used for identifying harmful data, ensuring effective detection of large-scale attacks in IoT ecosystems. Our findings indicate that these models enhance IoT security by deploying efficient intrusion detection systems (IDS) using reliable, large-scale datasets. The study highlights the performance of these models in balancing security and resource management, given the constraints of IoT devices.ML and DL approaches offer significant security benefits for IoT networks, despite the challenges associated with their implementation. The study underscores the importance of future research to address these challenges and further improve IoT security. The results provide valuable insights into the application of ML/DL models in IoT security, contributing to both theoretical knowledge and practical solutions for enhancing cybersecurity in IoT ecosystems.
A Comprehensive Review of Deep Learning Techniques for Intrusion Detection in the Internet of Medical Things Aisha Essa Muhammad; Amer Abdulmajeed Abdulrahman
Journal of Applied Engineering and Technological Science (JAETS) Vol. 7 No. 1 (2025): Journal of Applied Engineering and Technological Science (JAETS)
Publisher : Yayasan Riset dan Pengembangan Intelektual (YRPI)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.37385/jaets.v7i1.6637

Abstract

The work revisits the security issues of Internet of Medical Things (IoMT) platforms and provides a list of deep learning models used for intrusion detection. The study fills the salient gap in early detection of actual IoMT system intrusions for enhanced medical device and data security. A wide-ranging and systematic investigation of deep learning models, such as CNNs, LSTMs, and hybrid ones (GNNs and BiLSTMs) recently introduced was carried out. These were then analyzed against well-known benchmark datasets, such as ToN-IoT and IoT-Healthcare Security and WUSTL-EHMS-2020, to consider the quality of their detection work on cybersecurity threats for IoMT systems. The results indicated high accuracy in cyber threat detection, reaching even 100% accuracy. But the challenges are still how to decrease false positives and improve the real-time performance of the model on robustness and generalization when making real-world applications. The research is literature-based and aimed to provide some further updates on a secure IoMT framework by identifying recent studies in the security of the IoMT ecosystem and shedding light on future work using hybrid methods, blockchain technology, or federated learning approaches that can contribute to the detection of IDSs. And all can help pave the way for a more secure, privacy-protecting IoMT that safeguards extremely sensitive medical data. The research also enhances the model: utilizing 15+ deep-learning models to propose an IoMT-resistant architecture. This can promote participation in the theoretical research and practical security protocols in the IoMT context, thus drawing attention and comprehension from researchers and practitioners to enhance security protocols.