Reproduce “Efficient Phishing URL Detection Using Graph-based Machine Learning and Loopy Belief Propagation” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Wenye Guo, Qun Wang, Hao Yue, Haijian Sun, Rose Qingyang Hu

A phishing URL detection system incorporates network-level features alongside traditional URL string features within a graph-based machine learning framework utilizing Loopy Belief Propagation. The model, enhanced with a novel convergence strategy and refined edge potential, achieved F1 scores up to 98.77% on real-world datasets, demonstrating improved resilience and scalability.

Key insights

  • Integrating network-level features with URL string characteristics provides a more resilient feature set against adversarial phishing tactics.
  • Leveraging graph-based learning with LBP effectively models complex relationships between URLs and their associated features, enhancing detection by considering collective evidence.
  • A novel convergence strategy for LBP, involving dynamic cycle management, significantly improves inference stability and accuracy in graphs with hidden variables.
  • Dynamically adapting edge potentials based on entity similarity and label relationships refines the graph model's ability to differentiate between benign and malicious entities.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.