Harnessing XGBoost's Tuned Learning Rate: A Cloud Application for Superior Phishing URL Detection

Living in a time where most interactions are through the web, authenticity is everything. Among all the various security threats on the internet, phishing attacks—primarily through malicious URLs—stand at the front. Therefore, the paper introduces a proposed cloud-based application that would contribute to fighting such menaces by implementing the power of the XGBoost machine learning algorithm in practice. While XGBoost has achieved excellent results in classification tasks, mainly contributing to fine-tuning its learning rate to achieve the best performance. After applying RandomizedSearchCV, the viewpoint of optimal learning rate parameters improved Phishing detection accuracy systematically. It is used here as a cloud application, and scalable real-time phishing URL detection helps keep the internet safer for browsing. Preliminary results indicate the superiority of the tuned XGBoost model over traditional detection methods. This algorithm optimizes cybersecurity applications through machine learning.