A Systematic Approach for Malware URL Recognition
Nora Abdulrahman ALfouzan, C. Narmatha · 2022
In today's world, new communication platforms have had a profound effect on business promotion and growth over a broad range of applications, comprising social networking, e-commerce, and online banking. Regrettably, technological advancements are accompanied by new sophisticated tactics that target people in order to defraud them. Finally, the user's device can be compromised by identity fraud, revenue, or malware. In order to initiate attacks, a malicious Uniform Resource Locator (URL) or malicious website hosts a different unsolicited material in the structure of email, drive-by downloads, or phishing. Unsuspecting people access those websites and become victims of a variety of hoaxes, which includes the personal information theft (credit card, identities, and so on), financial ruin, and the installation of malware. The blacklist approach is the most popular technique utilized by many antivirus companies for detecting malignant URLs. Blacklists are simply the directory of URLs that have already been determined to be malicious. However, keeping a complete list of malicious URLs was nearly not possible, particularly because new URLs are created on a daily basis. Machine Learning methods utilize a collection of URLs as training data and develop a predictive algorithm based on mathematical properties to characterize a URL as malicious or not. Unlike blacklisting tactics, this enables them to popularize new URLs. This survey reviews the work related to the detection of malicious URL research using in the machine learning part.