Advancing Phishing Threat Mitigation: An Intelligent System for Optimized URL-Based Detection using AI

Phishing efforts are increasing lately due to the enhanced internet penetration worldwide. The purpose of this research is to design a phishing detection system that is both efficient and effective. This effort has acquired growing significance in boosting cybersecurity against the increasing threat of phishing assaults. The proposed approach is for data preparation and utilizing SMOTE for dataset balance, training/ testing on a dataset of 11,430 URLs with 87 features derived from URL structure, content, and external services. The findings reveal that CatBoost delivers the best accuracy, thereby performing better with fewer misclassifications than Random Forest, evidenced by its confusion matrix. The research reveals that integration of machine learning models especially CatBoost findings in phishing detection systems has enhanced its accuracy and reliability to 97.34%. Such improvement provides a robust approach to limiting phishing hazards.