Scholarlink

<p>The Internet of Things (IoT) ecosystem has revolutionized various industries but remains vulnerable to significant security challenges, including privacy breaches, network intrusions, and centralized vulnerabilities. To address these issues, this study introduces a Machine Learning-based Lightweight Blockchain Security Paradigm (ML-LBSP) that integrates blockchain technology with Agile Modified Dynamic Routing Optimization Capsule Neural Networks (AMDRO-CapsNets). The objective is to enhance IoT security by combining blockchain’s decentralized privacy-preserving capabilities with machine learning’s anomaly detection efficiency. The proposed framework employs a lightweight Proof of Stake (PoS) consensus mechanism to ensure energy-efficient and secure data validation, while AMDRO-CapsNets improve anomaly detection accuracy through dynamic routing optimization. Experimental validation was conducted using two benchmark datasets, CSE-CIC-IDS-2018 and N_BaIoT, which represent diverse IoT attack scenarios. The ML-LBSP framework achieved classification accuracies of 99.70% and 91.69%, respectively, significantly outperforming traditional methods. Additionally, the framework reduced execution time to 0.5&#xa0;s, demonstrating its suitability for resource-constrained IoT environments. These results highlight the potential of ML-LBSP to provide scalable, energy-efficient, and robust security solutions for IoT networks. In conclusion, the integration of blockchain and machine learning in the ML-LBSP framework offers a novel approach to addressing IoT security challenges, paving the way for future advancements in secure and efficient IoT systems.&#xa0;</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Enhancing cybersecurity through hybrid blockchain-enabled intrusion detection systems: A machine learning approach

  • 期刊论文

  • 发表日期：

    2025年8月20日

查看全文
  • Isaac Osei Asante,
  • Libing Wu

摘要

The Internet of Things (IoT) ecosystem has revolutionized various industries but remains vulnerable to significant security challenges, including privacy breaches, network intrusions, and centralized vulnerabilities. To address these issues, this study introduces a Machine Learning-based Lightweight Blockchain Security Paradigm (ML-LBSP) that integrates blockchain technology with Agile Modified Dynamic Routing Optimization Capsule Neural Networks (AMDRO-CapsNets). The objective is to enhance IoT security by combining blockchain’s decentralized privacy-preserving capabilities with machine learning’s anomaly detection efficiency. The proposed framework employs a lightweight Proof of Stake (PoS) consensus mechanism to ensure energy-efficient and secure data validation, while AMDRO-CapsNets improve anomaly detection accuracy through dynamic routing optimization. Experimental validation was conducted using two benchmark datasets, CSE-CIC-IDS-2018 and N_BaIoT, which represent diverse IoT attack scenarios. The ML-LBSP framework achieved classification accuracies of 99.70% and 91.69%, respectively, significantly outperforming traditional methods. Additionally, the framework reduced execution time to 0.5 s, demonstrating its suitability for resource-constrained IoT environments. These results highlight the potential of ML-LBSP to provide scalable, energy-efficient, and robust security solutions for IoT networks. In conclusion, the integration of blockchain and machine learning in the ML-LBSP framework offers a novel approach to addressing IoT security challenges, paving the way for future advancements in secure and efficient IoT systems.