Cyrillic URL: A Malicious Website Classification Using Deep Neural Network
摘要
In the contemporary period characterized by the proliferation of digital technology, individuals are consistently engaged with online resources. Web phishing is a form of social engineering attack in which individuals are deceived by cyber-attackers into divulging crucial login information. As the web evolves beyond English-speaking users, this sort of threat is increasing. A Cyrillic URL is a URL that includes a Cyrillic character with the intention of deceiving ordinary individuals. Cyber attackers transmit Cyrillic URLs by email or chat, prompting individuals to divulge their crucial passwords. This work presents a parsing of the deep web and the creation of a dataset consisting of 1 lakh benign and Cyrillic URLs. We have attained a classification accuracy of 97% by employing Bi-LSTM and have conducted a comparative analysis of different recurrent neural networks.