Integrated Machine Learning Approach to Phishing Detection: Comparing SVM, Random Forest, and XGBoost Models

The objective of this research is to create a phishing detection system that is both efficient and effective. This system will be based on sophisticated machine learning models, including SVM, Random Forest, and XGBoost. Given the growing threat phishing assaults pose, this research is vital for improving cybersecurity policies. The approach consists of data preparation, SMOTE for dataset balance, and training/ testing on 11,430 URLs with 87 characteristics obtained from URL structure, content, and other services. Based on its confusion matrix, which shows less misclassifications than Random Forest, results reveal that XGBoost performs with the best accuracy. Finally, including machine learning models-especially XGBoost-helps phishing detection systems to be much more accurate with 97.37% accuracy and reliability, therefore offering a strong means of reducing phishing risks.