Malicious URL website detection using ensemble machine learning approach

Phishing websites represent a vital cybersecurity threat that pretends to be reliable platforms to extract sensitive information from users. The detection of zero-day phishing attacks by blacklist-based filtering becomes challenging because these methods need regular updates from human operators. The proposed solution for this research depends on an ensemble machine learning framework using Random Forest and Decision Tree classifiers to extract features and classify phishing websites. The detection system identifies phishing sites through evaluation of URL patterns together with domain properties and site attributes. The model undergoes training with URLs obtained from authentic sources which contain both legitimate and phishing web pages. The project deploys a Flask web interface for phishing detection that provides real-time protection during security maintenance. Multiple assessments of the ensemble machine learning system demonstrate its better performance compared to standard detection methods for accuracy and real-time operations along with its adaptability. The research contributes to cybersecurity by delivering an automatic system which provides effective and scalable phishing detection capabilities.