PHISHCATCHER CLIENT SIDE DEFENCE AGAINST WEB SPOOFING ATTACKS USING MACHINE LEARNING

Abstract

The rapid growth of online banking, e-commerce, social media, and cloud services has increased the risk of phishing and web spoofing attacks. Fraudulent websites imitate legitimate websites to deceive users into revealing sensitive information such as passwords, financial details, and personal data. Traditional blacklist and signature-based methods often fail to detect newly created and zeroday phishing websites. To address this limitation, this project proposes PhishCatcher, a machine learning-based client-side system for detecting phishing and spoofed websites in real time. The system extracts URL-based, domain-based, HTTPS security, webpage-based, and lexical features. The dataset undergoes preprocessing, feature extraction, selection, normalization, and encoding before training. Random Forest, Decision Tree, Support Vector Machine (SVM), and XGBoost algorithms are evaluated using Accuracy, Precision, Recall, F1- Score, ROC-AUC, and Confusion Matrix. The bestperforming model is integrated into a browser protection mechanism to detect suspicious websites and alert users. The proposed system provides a lightweight, scalable, privacy-preserving, and intelligent approach for safer web browsing. Keywords: Phishing Detection, Web Spoofing, Machine Learning, Random Forest, SVM, XGBoost, Client-Side Security.