A review on malicious link detection techniques
Ashim Chaudhary, K.C. Krishna, Md Shadik, Dharm Raj · 2023
Recent advancement in technology has led to more and more use of the internet and its services. Many people use the internet for their day-to-day tasks like buying things from the internet, chatting with friends through social media, sending mail, and it is also used by many businesses and corporations. We all know that the Internet has made our connectivity more efficient and we are totally dependent on it. Though the internet has many advantages, one of its disadvantages can be taken as the privacy issue and theft of data via unauthorized access. There are many ways through which hackers or eavesdrops try to get people&s;s personal data and one of the ways is by sending or embedding malicious links or phishing links via mail attachments or embedding it on the website. So different ways have been developed to detect such malicious links. Their design includes features for a variety of harmful detection techniques, including blacklist approach, whitelist approach, visual-similarity based, content-based and mostly URL-based approach. Each has unique benefits and disadvantages. In this overview study, we mainly focus on machine learning-based strategies for detection of malicious URLs. In this study, we explore the various detection processes and approaches used by URL-based features in order to comprehend their structure. The performance based on various dataset&s;s combinations of URL characteristics is then examined. In order to encourage the development of improved URL-based phishing detection systems, we wrap up our research.