Scholarlink

Malicious URLs or websites are a threat to cybersecurity, as they have been appearing in search engine results more and more. Since many of the current detection techniques focus on specific types of attacks and often rely on browser add-ons that use blacklists, many rogue websites do not get noticed. We propose a novel service called Privacy-Preserving Safe Browsing (PPSB), which protects the privacy of the users and provides reliable identification of dangerous URLs. PPSB provides strong security guarantees as opposed to the traditional Safe Browsing services since it safeguards users’ search histories and encrypts private information to prevent analysts or service providers from accessing it from the outside. URLs are classified based on safety into three categories using machine learning models such as Support Vector Machine (SVM), Random Forest, and Decision Tree. The suggested approach adopts an anonymous aggregate analysis to understand internet activity while keeping private user information protected using AES encryption. The highest F-measure in the range of 60% proved that the framework was successful in identifying bad URLs while keeping the private data of users concealed. This study fills important holes in the current cybersecurity frameworks by demonstrating how PPSB may offer a complete solution for identifying dangerous websites while protecting user privacy and data security.

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 Malicious URL Detection with Privacy-Preserving Cryptographic Machine Learning Models

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • J. Kavitha,
  • S. Abishek,
  • G. Janani Sri,
  • S. S. Sanchana,
  • N. Vaishnavi

摘要

Malicious URLs or websites are a threat to cybersecurity, as they have been appearing in search engine results more and more. Since many of the current detection techniques focus on specific types of attacks and often rely on browser add-ons that use blacklists, many rogue websites do not get noticed. We propose a novel service called Privacy-Preserving Safe Browsing (PPSB), which protects the privacy of the users and provides reliable identification of dangerous URLs. PPSB provides strong security guarantees as opposed to the traditional Safe Browsing services since it safeguards users’ search histories and encrypts private information to prevent analysts or service providers from accessing it from the outside. URLs are classified based on safety into three categories using machine learning models such as Support Vector Machine (SVM), Random Forest, and Decision Tree. The suggested approach adopts an anonymous aggregate analysis to understand internet activity while keeping private user information protected using AES encryption. The highest F-measure in the range of 60% proved that the framework was successful in identifying bad URLs while keeping the private data of users concealed. This study fills important holes in the current cybersecurity frameworks by demonstrating how PPSB may offer a complete solution for identifying dangerous websites while protecting user privacy and data security.