Date of Publication :4th April 2024

Abstract:Phishing is a sort of online fraud in which fraudulent emails and websites deceive victims into disclosing important information. To combat this problem, scientists have developed tactics aimed at creating effective phishing URL detection systems. To this end, our method examines URL attributes such as domain-based and address-based data using feature engineering and ensemble machine learning techniques. This work presents an advanced phishing URL detection system that analyses URL properties using feature engineering and ensemble machine learning approaches. Our approach incorporates the Gradient Boosting Classifier, outperforming other algorithms that mostly use Random Forest, Decision Tree, and Logistic Regression, and achieves an amazing 97.6% accuracy in differentiating between phishing and authentic websites. Phishing is a common online danger that uses shady websites and emails to trick people into disclosing personal information. Our solution guarantees accuracy and efficiency while addressing issues related to computational complexity. We showcase our system's performance using precision, recall, accuracy, and F1 score metrics through a thorough examination of various datasets. Furthermore, our system has an intuitive user interface that combines HTML, CSS, and Flask to improve usability and accessibility. By making a substantial improvement to phishing detection, this study protects consumers from online risks and maintains their privacy

Reference :