Abstract

Phishing attacks pose a significant threat to online security by tricking users into revealing sensitive information through deceptive websites. These attacks are frequently used as entry points for broader cyber intrusions, risking both personal and organizational data. Traditional detection systems often rely on predefined rules or manual feature extraction, which limits their effectiveness, especially against new or zero-day phishing threats. To overcome these limitations, this project proposes an intelligent phishing website detection system utilizing an ensemble machine learning approach. By combining the strengths of multiple learning algorithms, the model enhances accuracy and adaptability. The system automatically extracts meaningful features from URLs and web page content, enabling it to distinguish between legitimate and malicious websites efficiently. This approach not only improves the detection rate but also reduces false positives, making it suitable for real-world deployment. The ensemble strategy ensures the model remains resilient to evolving attack patterns, offering a more robust and scalable solution to phishing detection.