Scholarlink

<p>This research focuses on detecting malicious URLs using machine learning methods. We used supervised machine learning models to distinguish between malicious and benign URLs, experimenting with several algorithms including Logistic Regression, SVM, Decision Tree, Random Forest, and Gradient Boosting. The URLs were analyzed using 50 features categorized into five classifications: lexical-based, domain-based, content-based, anomaly-based, and HTML/JS-based features, which helped improve the models’ accuracy. Key findings show that the Gradient Boosting algorithm demonstrated superior performance with the highest accuracy of 99.81% compared to other algorithms. The research highlights the increasing importance of accurate malicious URL detection methods in today’s context of growing internet usage and associated cyber threats. We propose that detection capabilities can be further enhanced by incorporating Mongolian URLs into the dataset and adding additional content-based features.</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

Detecting Malicious URLs Using Machine Learning

  • 期刊论文

  • 发表日期：

    2025年6月16日

查看全文
  • Munkhbayar Bat-Erdene,
  • Oyuntungalag Buyantur,
  • Densmaa Batbayar,
  • Byambadorj Dondogmegd,
  • Juhee Choi

摘要

This research focuses on detecting malicious URLs using machine learning methods. We used supervised machine learning models to distinguish between malicious and benign URLs, experimenting with several algorithms including Logistic Regression, SVM, Decision Tree, Random Forest, and Gradient Boosting. The URLs were analyzed using 50 features categorized into five classifications: lexical-based, domain-based, content-based, anomaly-based, and HTML/JS-based features, which helped improve the models’ accuracy. Key findings show that the Gradient Boosting algorithm demonstrated superior performance with the highest accuracy of 99.81% compared to other algorithms. The research highlights the increasing importance of accurate malicious URL detection methods in today’s context of growing internet usage and associated cyber threats. We propose that detection capabilities can be further enhanced by incorporating Mongolian URLs into the dataset and adding additional content-based features.