A Machine Learning Approach for Identifying and Detecting Hazards

Abstract

Every day, we log on to the internet and use it to conduct our business. As a result, browser vendors compete for users' attention by adding new features and enhancing existing ones, which exposes websites to hackers' attacks. Surfers, on the other hand, are looking for a rapid and precise model that can tell the difference between safe and harmful websites. In this study, we develop a new classification approach to evaluate and detect dangerous online sites using Machine Learning classifiers such as Random Forest, Support Vector Machine, Naive Bayes, Logistic Regression, and a specific URL (Uniform Resource Locator). Experimental data shows that the random forest classifier is more accurate than other machine learning classifiers, with an accuracy rate of 95%.