An Effective Way to Detect Phishing Website Using Machine Learning

In today's interconnected digital landscape, phishing represents a significant threat to both individuals and organizations. Phishers cleverly utilize deceptive techniques to extract sensitive information from unsuspecting users, posing a risk to critical account details. Despite various proposed methods to detect phishing websites, the adaptability of phishers continually challenges the efficacy of existing approaches. Recognizing the vital role of Machine Learning (ML) in combating these evolving threats, this research investigates the application of ML techniques. Common characteristics in phishing attacks can be identified through accurate ML analysis. This research employs a range of ML models and transformer models, including Catboost (Cat), XGBoost (XGB), Random Forest (RF), Support Vector Machine (SVM), Decision Tree (DT), TabNet, K-nearest neighbor (KNN), and Logistic Regression (LR) with grid-search. Among these models, the Catboost classifier achieved a remarkable accuracy of 97%. This study shows dedication to utilizing diverse methodologies to advance the field of phishing website detection.