Detection of phishing URLs: a machine learning approach

Abstract

Phishing attacks represent one of the most prominent cybercrimes today, their main aim is illicitly acquiring sensitive information such as passwords, user names, email bank details, and credit card information. Their impacts extend across various sectors, including online payment platforms, financial institutions, and cloud storage providers. Usually, phishing attacks target websites associated with online payments and webmail services. Various techniques have been used to combat phishing attacks, including blacklisting, heuristic analysis, visual similarity checks, and machine learning. while blacklisting is usually used to ease implementation, it falls short in detecting new phishing attacks. Machine learning emerged as a highly efficient technique for detecting phishing attacks, comprehensively addressing the limitations of other methods. This research focuses on machine learning algorithms, namely logistic regression, decision trees, random forests, and support vector machines (SVM)for phishing URL detection.