A Systematic Review on Phishing Attacks Detection Techniques based on Machine Learning

Phishing attacks are a substantial cyber security threat because they use spoof e-mails and fake websites to target individuals, governments, and companies to steal critical information. Traditional detection methods, such as rule-based and blocklists, are losing effectiveness as phishing e-mail thr...

Full description

Saved in:
Bibliographic Details
Published in2025 3rd International Conference on Self Sustainable Artificial Intelligence Systems (ICSSAS) pp. 930 - 937
Main Authors Vikash, Sharma, Anuj Kumar
Format Conference Proceeding
LanguageEnglish
Published IEEE 11.06.2025
Subjects
Companies
Computer crime
cyber security
Deep learning
deep learning (DL)
Electronic mail
Government
machine learning (ML)
Performance metrics
Phishing
phishing attack
Security
Systematic literature review
Uniform resource locators
Online AccessGet full text
DOI10.1109/ICSSAS66150.2025.11080759

Cover

More Information
Summary:Phishing attacks are a substantial cyber security threat because they use spoof e-mails and fake websites to target individuals, governments, and companies to steal critical information. Traditional detection methods, such as rule-based and blocklists, are losing effectiveness as phishing e-mail threats become more sophisticated. These days, cybercriminals use trusted e-mail services and advanced strategies to get around security measures. To enhance detection, researchers utilize deep learning (DL) and machine learning (ML) more and more. This paper reviews more than 50 papers on ML and DL-based phishing detection techniques. It examines accuracy, F1 score, and other performance metrics of different ML and DL models on phishing e-mails, websites, and URL detection. It also compares different studies and highlights the weaknesses of every method. It explores the current challenges in phishing detection and proposes future research directions to improve its effectiveness.
DOI:10.1109/ICSSAS66150.2025.11080759