International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 09 Issue: 06 | June 2022
p-ISSN: 2395-0072
www.irjet.net
Phishing Detection using Decision Tree Model Aman Ahamed1, Dr. Ramananda Mallya K2, Anushri A Shetty3, Delisha DSouza4, Ashokkumar Tirumala Gopi5 1,3,4,5 Dept. of
Information Science and Engineering, Mangalore Institute of Technology & Engineering, Moodbidri. Associate Professor, Dept. of Information Science and Engineering, Mangalore Institute of Technology & Engineering, Moodbidri. ---------------------------------------------------------------------***--------------------------------------------------------------------2
Abstract - In the modern days the security is the main
based mailing systems in and online payments. The remaining 22 % of the attacks are made for industrial sectors.
concern in this rapidly evolving world with the technology advancement. There are many of the cases which led to huge number of financial losses by common social attacks. These attacks are the one that made technically or to the targeted device. It's in the form of the virus or Trojan or it may be in the form of a normal website link which we also called as the URL (Uniform Resource Locator).These URLs contains the software or the malicious program which takes out the users all the valuable and more secured and private information (or sensitive data) when this URL is entered by the user in his remote machine. This form of attack is known as Phishing. Normally the user will see the web page appearing as a simple and interactive but in behind it is more and more dangerous one. A fraudulent try made by the attacker in order to steal the users data all the private information like we have username, password, and private details like users financial bank account and details of the users credit card. To avoid these attacks there are many advancements in artificial intelligence and machine learning, which have efficient and more compact techniques to find out the fake URLs. A machine learning model made up of decision tree algorithm is developed which will scan and filtes out the common words and learns the specific features and then it will provide the appropriate result.
The consequences and the results when phishing attacks occur will cause huge financial losses in the case of the banking domain. The current era internet revolution has increasing and the advancement in technologies is also increasingly growing, it has become an attractive place for all potential users. Phishing is normally imitated by mimicking as a trustworthy person or an entity on the Internet which is done by integrating both social engineering and technological tricks. Lastly, we know that economic and financial helpers such as banks are now becoming more important on the Internet thereby making people's lives in this world easy. Security and the safety of the people against these frauds are mandatory in this digital era. Phishing is a major attack or threat when it comes to securing the website. There are mainly two types of phishing attacks one is called the Spear phishing, which means targeting the specific and private/public companies and the individual people. The other one is called Clone phishing. This means that this is an attack where the real or the original mail containing an additional attachment or the URL/link is copied to a fresh (new) mail with malicious attachment or URL [2].
Key Words: Uniform Resource Locator, Decision Tree, Security, Machine Learning
1. INTRODUCTION
2. BACKGROUND
Phishing in layman's terms is just giving the user by an attacker the web link or we say it's a programmed URL or abbreviated as Uniform Resource Locator where the term programmed contains the scripts or the virus or malicious infinite time running program or a zombie the process that when invoked runs itself and it will do those tasks or the commands ordered by the attacker.
The main goal to achieve successful phishing is the user's data, assets, or private information that is stolen through a fake website [3]. If we detect bad URLs in the early stage this is the best strategy to avoid contact with phishing websites. Phishing websites are to be determined through their basic domains [4]. These are related to the URL that needs to be registered. We will implement machine learning algorithms to classify the data in this case. The basic algorithms used here are as follows. The proposed technique gives 95% accuracy. This mainly depends on the quantity of data set divided into training and testing.
This URL seems to be the normal one. But the attacker uses this in order to get all the private and confidential information from the user so that there is some benefit enjoyed by the attacker. The domains are more. These attacks majorly occur in the field of online payment sector, web-based email, and in the cases of cloud storage [1]. 78 % of the attacks are made only in the domains like web-
© 2022, IRJET
|
Impact Factor value: 7.529
|
ISO 9001:2008 Certified Journal
|
Page 2458