A PHISHING DETECTION SYSTEM BASED ON INTELLIGENT DEEP LEARNING TECHNIQUES

Abstract

In contrast to software vulnerabilities, phishing websites or URLs are a type of internet security problem that focuses on human vulnerabilities. There are many ways to compromise an internet user's security, but the most popular one is phishing, an attack that seeks to obtain or misuse a user's personal data, such as passwords, credit card information, identity, and account information. Phishers collect information about users by impersonating legitimate websites that are indistinguishable to the naked eye. Users' sensitive information may be accessed, putting them at risk of financial harm or identity theft. As a result, there is an urgent need to create a system that effectively detects phishing websites. This paper proposes three distinct deep learning-based techniques for detecting phishing websites, including long short-term memory (LSTM) and convolutional neural network (CNN) for comparison, and finally an LSTM-CNN-based approach. Experimental findings demonstrate the accuracy of the suggested techniques, i.e., 99.2%, 97.6%, and 96.8% for CNN, LSTM–CNN, and LSTM, respectively. The proposed phishing detection method demonstrated by the CNN-based system is superior.