Malicious URLs Detection Using Machine Learning Models

This research focuses on developing a machine learningbased system for determining malicious links.The main goal is to distinguish between benign and malicious web links using automated classification techniques.To achieve this, we used multiple datasets, including the Malicious URLs Dataset by Manu Siddhartha and the URL Dataset by Antony J.These datasets provide a vast compilation of URLs labeled in different categories, making it easy for effective model training and evaluation.This approach involves training various machine learning algorithms, comprising Random Forest, Gradient Boosting, and Multi -Layer Perceptron models.Feature extraction procedures such as onehot encoding and URL parsing etc. were put into practice to maximize model's performance.The trained models achieved an accuracy of 85%, and have shown substantial possibilities for realworld cybersecurity applications.This prototype serves as a promising move toward automated URL threat detection, with future work focusing on improving detection efficiency and managing evolving cyber threats.