Phishing Attack Response and Risk Mitigation Using Deep Learning and Machine Learning
Usharani Bhimavarapu · 2025
Phishing attacks have emerged as one of the most prevalent and damaging forms of cyber threats, posing significant risks to the security and privacy of users on the internet. These attacks typically involve fraudulent websites designed to deceive users into disclosing sensitive information, such as passwords, credit card details, and personal data. Detecting phishing websites has become a critical task in cybersecurity, as traditional methods often fail to effectively identify new and sophisticated phishing tactics. This study addresses this challenge by leveraging a carefully curated dataset containing 10,000 instances of URLs and web page features, specifically designed for phishing detection. A combination of advanced machine learning techniques, including feature extraction using ResNet50 and classification with Support Vector Machine (SVM), is employed to accurately distinguish between legitimate and phishing websites. SMOTE is used for data balancing to mitigate class imbalance, ensuring that both classes are adequately represented during training. The proposed approach demonstrates a significant improvement in detecting phishing websites by effectively capturing complex patterns and characteristics indicative of phishing activity. The results underscore the potential of deep learning and machine learning methods in enhancing the accuracy and reliability of phishing detection systems.