DeMalFier: Detection of Malicious web pages using an effective classifier

Abstract

The web has become an indispensable global platform that glues together daily communication, sharing, trading, collaboration and service delivery. Web users often store and manage critical information that attracts cybercriminals who misuse the web and the internet to exploit vulnerabilities for illegitimate benefits. Malicious web pages are transpiring threatening issue over the internet because of the notoriety and their capability to influence. Detecting and analyzing them is very costly because of their qualities and intricacies. The complexities of attacks are increasing day by day because the attackers are using blended approaches of various existing attacking techniques. In this paper, a model DeMalFier (Detection of Malicious Web Pages using an Effective ClassiFier) has been developed to apply supervised learning approaches to identify malicious web pages relevant to malware distribution, phishing, drive-by-download and injection by extracting the content of web pages, URL-based features and features based on host information. Experimental evaluation of DeMalFier model achieved 99.9% accuracy recommending the impact of our approach for real-life deployment.