Web-Based Machine Learning Framework for Phishing URL Detection and Analysis

Phishing is a cyber attack in which malicious actors attempt to deceive users into providing sensitive information, such as login credentials, financial details, or personal identification. These attacks are primarily carried out through fraudulent URLs that direct unsuspecting victims to websites designed to look like trusted platforms. In this research paper, we introduce a web-based ML framework for performing PUD. The proposed framework not only classifies URLs but also analyzes embedded links and performs domain validation using WHOIS data. Results indicate efficient performance for the framework in the classification tasks with <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$A$</tex> of 0.96875, <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$P$</tex> of 0.94 and 1.00 for bad and good, respectively, <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$R$</tex> of 1.00 and 0.94 for bad and good, respectively, and <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$F$</tex> of 0.97 and 0.97 for bad and good, respectively. Additionally, results also highlight results for the WHOIS analysis, and how legitimate and phishing URLs can parsed through the framework.