10.55041/IJSREM43379

Phishing websites pose a significant cybersecurity threat, deceiving users into disclosing sensitive information such as login credentials, financial details, and personal data. Traditional rule-based detection methods struggle to keep up with evolving phishing techniques, necessitating more intelligent and adaptive solutions. This paper explores the application of machine learning (ML) for phishing website detection, leveraging URL-based, domain-based, and content-based features. Various ML algorithms, including XGBoost, Support Vector Machines (SVM), Random Forest, and Neural Networks, are evaluated for their effectiveness in identifying phishing websites. The study also addresses key challenges such as dataset imbalance, feature selection, adversarial attacks, and the need for real-time detection. Experimental results demonstrate the effectiveness of ML-based models in improving detection accuracy and reducing false positives. The paper concludes with recommendations for enhancing phishing detection systems through feature engineer- ing, ensemble learning, and deep learning techniques. Index Terms—Phishing Detection, Machine Learning, Cyberse- curity, URL Analysis, Feature Engineering, Adversarial Attacks.