A Machine Learning Model to Detect Phishing Emails using Ensemble Technique
Loading...
Date
2024-11
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Kenyatta University
Abstract
The majority of phishing attacks prey on behavioral flaws in users. Phishing links are included in an email that an attacker sends to the recipient that looks and feels authentic. The attacker can obtain sensitive data, like as usernames, passwords, and credit card details, by having the receiver click on the embedded links and access the hacked account. With the increasing case of cyber-attacks, organizations are looking for safer ways of protecting data and preventing getting hacked or getting hacked again. Design and technology should be greatly improved to prevent hackers from infiltrating networks. Phishing attacks, which mostly target financial organizations, have been identified as the most common online content attack according to surveys. A 2017 Ponemon Institute LLC survey estimated that the yearly loss from phishing attempts is almost $1.5 billion. The Internet of Things (IoT) is contributing to the global danger to information security; hence, a more effective phishing detection system is needed to reduce these losses and reputation injury. In order to increase the accuracy of phishing detection and prevention, this research study investigates and reports the use of several machine-learning models by utilizing more phishing email features and the random forest algorithm. To detect and prevent phishing attacks, this project examined current phishing techniques, examined the impact of using an ensemble model, designed and created a supervised classifier to identify and stop phishing emails, and tested the model using available data. The model was learned under supervision using a dataset of legitimate and fraudulent emails. With a rate of less than 0.1% for False Positives (FP) and False Negatives (FN), the expected accuracy is 99.9% which will be higher than the already existing models therefore better detection of fraudulent emails.
Description
A Research Project Report Submitted In Partial Fulfilment of the Requirements for the Award of the Degree of Masters of Science in Computer Science in the School of Pure and Applied Sciences of Kenyatta University, June 2024.
Supervisors
Abraham Mutua
Stephen Waithaka