Detection of Malicious URLs

Abstract

AbstractIn today’s world, the usage of the web is increasing for skilled and private use. Most of our activities are associated with the online, for example, surf-riding online for few data, sending or receiving mails, etc. Web malware attacks and its area units are susceptible to most of the vital sectors which ends in significant loss in terms of wealth and security. Hackers host uninvited content and attract naive users to become victims of scams, like stealing non-public data and malware installation. Malware attacks result in revenue losses in businesses. Malicious URLs play a major role in cyberattacks these days. Malicious URLs may be delivered to a person through messages, emails, or advertisements. It is imperative to act on such threats in a very timely manner. The traditional way to detect malicious URLs is by blacklisting which is done by the use of web databases. The most important disadvantage of blacklisting is that it fails to find the new URLs. The projected approach identifies Uniform Resource Locators (URL) by employing a machine learning approach of feature extraction, and its algorithmic rule called as long short-term memory (LSTM) and predicts whether it is attacked (malicious) URLs or benign URLs.KeywordsBenignCyber threatsMaliciousMachine learningUniform Resource LocatorLong short-term memory