EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

A Robust Intelligent Approach for Phishing URL Detection Using Hybrid Machine Learning

Hind Abdulkareem Abdalrazaq

CyberSystem Journal · 2025

DOI: 10.57238/csj.2025.1017 ↗

Kurz gefasst

Fachartikel in der Zeitschrift CyberSystem Journal, 2025. Verfasst von Hind Abdulkareem Abdalrazaq. Die Arbeit nennt 42 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Phishing attacks are still the main and most severe vulnerability to cyber security through deceptive URLs for sensitive user information. Therefore, this paper proposes a framework based on deep learning techniques for phishing and legitimate URLs as an alternative to the currently employed blacklist-based detection approaches that have proven rather limited. The architecture includes one custom Convolutional Neural Network (CNN) and three Transfer Learning Architectures, ResNet50, InceptionV3, and VGG16 using representations of features based on URLs. All models under consideration shall be trained with Adam Optimizer and Binary Cross-Entropy loss function so that a fair comparison can be made under unified experimental set-up conditions. The set is broken down as 70% training, 10% validation, and 20% testing. Experimental results provide clear evidence that transfer learning models perform much better than the baseline CNN. The InceptionV3 model posted a validation accuracy of 100% leading the pack, followed by VGG16 at 98.96%. ResNet50 could muster only 97.92%. The proposed CNN model has also achieved quite competitive performance with a validation accuracy above 97%. Therefore, these results are explicit in confirming the effectiveness, robustness, and generalization capability of deep learning architectures toward the problem of phishing URL detection. This proposed framework will go a long way in ensuring web security is beefed up as it creates a barrier against all evolving cyber threats.

Zitieren

APA

Abdalrazaq, H. A. (2025). A Robust Intelligent Approach for Phishing URL Detection Using Hybrid Machine Learning. CyberSystem Journal. https://doi.org/10.57238/csj.2025.1017

Vancouver

Abdalrazaq HA. A Robust Intelligent Approach for Phishing URL Detection Using Hybrid Machine Learning. CyberSystem Journal. 2025. doi:10.57238/csj.2025.1017

BibTeX
@article{Abdalrazaq2025,
  author = {Abdalrazaq, Hind Abdulkareem},
  title = {A Robust Intelligent Approach for Phishing URL Detection Using Hybrid Machine Learning},
  journal = {CyberSystem Journal},
  year = {2025},
  doi = {10.57238/csj.2025.1017}
}
Kostenlos starten →