Leveraging Machine Learning to Detect Phishing Websites: Protecting Individuals and Businesses from Cyberattacks

Phishing attacks are one of the emergent threats in the cybersecurity domain as it involves malicious websites and emails focused on compromising the sensitive information of the victim. The research paper is focused on detecting the phishing URLs with the help of machine learning. The dataset contains more than 11,000 URLs; every URL is labeled as phishing or legitimate, with 89 features describing the URL, such as length, use of special characters, and domain details. The URLs are classified using ML models such as RF, SVC, and GBM. In this paper, a hybrid model is built by taking in an RF, an SVC, and a GBM, optimized using cross-validation and hyperparameter tuning. The efficiency of the models is grasped based on the evaluation metrics like precision, accuracy, recall, and F1-score. The hybrid model outperformed the individual algorithms with a detection accuracy of 95% on phishing URLs, thus indicating its ability to be used as an efficient phishing solution to enhance cybersecurity defenses.