Machine Learning-Powered Malicious Website Detection System

The internet's expansion has introduced significant cybersecurity risks, particularly through malicious websites. Traditional detection methods are often inadequate due to the complexity and volume of threats. This paper explores the use of machine learning (ML) techniques for identifying harmful URLs. Various models, including decision trees, logistic regression, support vector machines (SVM), and naive bayes, are evaluated using a dataset with features such as IP repute and domain age. Our results indicate that decision trees achieve the highest accuracy and F1-score for differentiating between benign and malicious URLs. This research underscores the potential of ML to improve cybersecurity and lays the groundwork for future advancements in malicious URL detection.