Hacker Eye

EasyChair Preprint 15938

Abstract

Phishing sites are a dangerous cyber threat where users are tricked into revealing sensitive information. This paper proposes Hacker Eye, a machine learning-based phishing prediction tool. The paper discusses various classification techniques, such as Random Forest, XG Boost, and ensemble techniques, to ensure maximum detection accuracy. Feature extraction techniques, such as URL structure inspection, domain age, and SSL certificate verification, assist in making classification more precise. A new hybrid technique involving deep learning and natural language processing (NLP) is also proposed to achieve maximum phishing detection. Test results indicate that ensemble techniques, such as XG Boost and Voting Classifiers, provide maximum accuracy in phishing and genuine site detection. The paper concludes with the description of how machine learning can be utilized to achieve maximum cyber defense and recommends future improvements to adaptive learning and multidisciplinary approaches.

Keyphrases: Cybersecurity, NLP, Phishing Detection, Random Forest, Website Classification, XG Boost, adaptive learning, deep learning, machine learning

@booklet{EasyChair:15938,
  author    = {M Suthar Kunta and Naga Surya Pakalapati and Priya Banik and V L S Sai Sathwik Badam and G S Navaneeth Arumilli},
  title     = {Hacker Eye},
  howpublished = {EasyChair Preprint 15938},
  year      = {EasyChair, 2025}}