ENHANCED MODEL FOR DETECTION OF PHISHING URL USING MACHINE LEARNING
Siddhesh Masurkar, Vipul Dalal · Ethics and Information Technology · 2020
Due to the increase in the use of the internet and the advancement of technology, more people are browsing on the internet.No need to mention that most people are not well aware of basic precautions, preventive measures that have to be taken while performing transactions over the net.This lack of knowledge among most people is well exploited by the phishers or attackers to acquire valuable private information from the user.Those security threats can cause user's mobile or computer to be hacked by the remote person usually called a hacker by executing some malware in the computer or mobile when the user clicks on the links purposely sent to the user.The phisher or Attacker or Hackers use sophisticated technics to cheat the user or to make scam in the virtual world by creating a fake site which looks almost similar to the original one.Therefore, if proper attention is not provided while filing user credentials or clicking on unknown links, those the security of the user can be compromised and its private information can be sent to the person who can take advantage of it.Phishing is new kind of an attack which is prevalent nowadays on the internet to jeopardize privacy of the user and subsequently of an organization.Therefore, many works are going on to detect and identify the phishing sites so user can be saved from entering his personal information and to prevent security attacks.But at the same time the way of designing phishing links are also evolving by improving content of the link in such way that the detection rate would below.This paper is presenting a detection system which give emphasis on boosting algorithm.At the same time various other classifiers are used on same set of datasets to evaluate the performance of the Boosting algorithm.We implemented five classification algorithms and some unusual features from the URLs.The experimental results exhibit better accuracy and other parameters of the Boosting algorithm as compared to other traditional classifiers.