Abstract

Malicious websites hosting drive-by download exploits have become more common as the internet has increased in popularity. To facilitate malware, attackers are increasingly redirecting users away from ordinary websites to attack pages. In the interim, the page''''''''s CSS settings are being tweaked to prevent modifications to the aesthetic impacts of specific pages. According to the researchers, positive examples include malicious drive-by-download web pages, while negative instances include benign web pages. Malign pages are discussed in the unsecured work and spam. To detect malign pages from pages of any website, first, manually crawl a list of safe URLs to collect the data. Furthermore, ensemble classifiers are used in work to train and test the model, i.e., Random Forest, Logistic Regression, Naive Bayes, and lastly, an ensemble of all the models is used. At last, a comparison is shown based on the accuracy score of different classifiers.