Designing and Developing PHISHR: A Machine Learning-Based System for Phishing Attack Domain Detection

Abstract

Phishing attacks, which take advantage of people's weaknesses to trick them into disclosing critical information, are still a severe danger to cyber security. This paper suggests a machine learning-based method for identifying phishing attack domains. Using supervised learning techniques, our algorithm uses a wide range of information from website content, domain names, and historical data to categorize domains as phishing or authentic. We use a dataset that includes tagged examples of valid domains and phishing attempts to train and assess our model. Our system successfully distinguishes malicious and legitimate domains with robust detection performance through feature selection and ensemble learning techniques. Our technology improves proactive defence mechanisms against phishing attempts by automating the detection process, strengthening the overall cyber security posture. The creation of a thorough phishing detection system that combines Fast API, machine learning techniques, and a Chrome extension is presented in this work. The system aims to guarantee scalability and user-friendliness while offering real-time defence against phishing attempts. Machine learning algorithms analyze different aspects of domain names and correctly identify phishing and legal websites. Rapid API allows the Chrome extension and backend services to communicate seamlessly, facilitating effective data processing and response creation. The Chrome extension provides the user interface, making it easy for users to see possible phishing websites while they're online. By taking a comprehensive approach, the solution provides users with proactive defence mechanisms against phishing threats constantly developing in the digital realm.