URL Based Malicious Activity Detection Using Machine Learning

Abstract

The constant use of the Internet has led to an increased vulnerability to malware attacks through malicious websites. The goal of this research is to create a machine-learning algorithm that will detect whether URLs contain susceptible activities such as viruses, phishing, malware, worms, etc. or are secure. Malicious URLs are compromised URLs that are employed in drive-by downloads and online attacks. Phishing and social engineering are common types of attacks that use malicious URLs. The fact that one-third of all websites have the potential to be harmful shows how widespread bad URLs are in online crime. This work deals with three machine learning models, such as random forest, light GBM, and XG Boost, to analyse our data and give the best one as per the results and analysis.