Phishing Attack Detection Using Convolutional Neural Networks

Abstract

Phishing attacks are a prevalent form of social engineering that target individuals through emails to obtain confidential and sensitive information. These attacks can lead to larger security breaches in both corporate and government networks. There have been several attempts to counter phishing assaults, but so far none have proven successful. For this reason, improved strategies for identifying phishing attempts are desperately needed. The proposed fix is a deep learning-based strategy for identifying malicious phishing attempts. By analyzing more than 5,000 phishing emails sent at the University of Malaysia’s Department of Computer Science and Information Technology, the authors hoped to create a model that reliably detects phishing assaults to achieve this, they selected relevant features through feature engineering and used the Random Forest models to extract feature importance at different levels. Finally, the model was trained using Convolutional Neural Networks (CNN), leading to improved detection and accuracy.