Proposing a Measure for Computing Security Risk of Links to Prevent Internet Frauds

Malicious links are sent by intruders in various ways such as email, SMS, or social networks for the purpose of fraud or abuse. Various methods have been presented so far to identify malicious links, all of which have errors. On the other hand, by identifying how these methods work, hackers have changed their techniques and are able to present their malicious link as a useful link to the victim. The purpose of this article is to present a new method to identify and deal with malicious links. For this purpose, after reviewing the previously presented researches in the field of identifying malicious links and security risk criteria, the concept of security risk for links has been redefined and a new criterion has been presented to calculate it. This measure uses the nearest neighbor method to estimate the security risk. Based on these criteria, a security score can be calculated for each provided link so that the user can decide whether or not to use the link, which will reduce internet fraud. Experiments performed on two real datasets related to previously known malicious and useful links show the effectiveness of the proposed criterion in terms of detection rate. That is, this criterion is able to identify a high percentage of malicious links in the datasets.