Scholarlink

<p>Cybercriminals have been exploiting online learning platforms since most educational institutions adopted distance learning, notably during the COVID-19 pandemic, which led to the growth of e-learning. This development increases cyberattacks, as students have become targets due to the sensitivity of their information. Although current methods, such as signature-based filters, encryption methods, intrusion detection systems, and conventional machine learning classifiers, provide security protection, they have high false-positive rates, high computational cost, and low adaptability. The Convolutional Neural Network-Orca Optimization Algorithm (CNN-OOA) is a hybrid method that is used to identify cyber threats in online learning. We prioritize phishing threats as they are among the most prevalent and damaging types of cyberattacks in the e-learning ecosystem, where stolen credentials are used to breach student confidentiality via emails or websites that appear to be legitimate, thereby compromising sensitive personal information such as login credentials and financial data. The key novelty of our approach is the use of OOA’s metaheuristic optimization to tune the CNN hyperparameters in real time. This overcomes static models and enhances the detection rates for previously unknown phishing attacks. The results of comprehensive experiments show that CNN-OOA achieved 96.89% accuracy, 98.17% precision, 95.68% recall, 96.91% F1-score, and 99.49% AUC on phishing emails, and 98.04% accuracy, 98.60% precision, 97.10% recall, 97.85% F1-score, and 99.60% AUC on phishing e-learning websites. These findings indicate that the proposed deep learning approach effectively enhances the cybersecurity of e-learning platforms.&#xa0;The full implementation and associated resources are publicly available at:<a href="https://doi.org/10.5281/zenodo.18566604">https://doi.org/10.5281/zenodo.18566604</a>.&#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

CNN-OOA-Based Cyber Threat Detection: Protecting E-Learning from Phishing

  • 期刊论文

  • 发表日期：

    2026年4月5日

查看全文
  • Brahim Khalil Sedraoui,
  • Abdelmadjid Benmachiche,
  • Amina Makhlouf,
  • Khadija Rais,
  • Chaouki Chemam

摘要

Cybercriminals have been exploiting online learning platforms since most educational institutions adopted distance learning, notably during the COVID-19 pandemic, which led to the growth of e-learning. This development increases cyberattacks, as students have become targets due to the sensitivity of their information. Although current methods, such as signature-based filters, encryption methods, intrusion detection systems, and conventional machine learning classifiers, provide security protection, they have high false-positive rates, high computational cost, and low adaptability. The Convolutional Neural Network-Orca Optimization Algorithm (CNN-OOA) is a hybrid method that is used to identify cyber threats in online learning. We prioritize phishing threats as they are among the most prevalent and damaging types of cyberattacks in the e-learning ecosystem, where stolen credentials are used to breach student confidentiality via emails or websites that appear to be legitimate, thereby compromising sensitive personal information such as login credentials and financial data. The key novelty of our approach is the use of OOA’s metaheuristic optimization to tune the CNN hyperparameters in real time. This overcomes static models and enhances the detection rates for previously unknown phishing attacks. The results of comprehensive experiments show that CNN-OOA achieved 96.89% accuracy, 98.17% precision, 95.68% recall, 96.91% F1-score, and 99.49% AUC on phishing emails, and 98.04% accuracy, 98.60% precision, 97.10% recall, 97.85% F1-score, and 99.60% AUC on phishing e-learning websites. These findings indicate that the proposed deep learning approach effectively enhances the cybersecurity of e-learning platforms. The full implementation and associated resources are publicly available at:https://doi.org/10.5281/zenodo.18566604.