Model Proposal for Phishing Detection in Text Messages Using Machine Learning

Sandoval J.; Ormeño V.; Ticona W. · Lecture Notes in Networks and Systems (2024) · DOI: 10.1007/978-3-031-70518-2_50

Resumen

In the dynamic landscape of digital communication, text messages are seamlessly integrated into everyday interactions, but they have become a prime target for cybercriminals using phishing attacks. With around 55% of users falling victim to such fraudulent tactics, addressing the growing threat of SMS phishing is imperative. This article presents an in-depth analysis of the central role of machine learning in cyber security, focusing on the detection and prevention of SMS phishing. Based on several works such as Moncada’s methodology, Dueñas’ study of AI applications, ALMahadin’s optimization based SVM model, Sheik’s email classification and Pandiyan’s algorithms to distinguish between legitimate and malicious websites, the study presents a robust methodology. This approach includes data collection, preprocessing, feature extraction using CountVectorizer and Tf-Idf, and data classification using various machine learning algorithms. The results show that the Random Forest Classifier is the optimal model, achieving an excellent accuracy of 97.94% in distinguishing between legitimate and potentially malicious text messages. The study concludes by highlighting the invaluable contribution of machine learning techniques in strengthening cyber security measures against the threat of SMS phishing. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.

Instituciones: Universidad ESAN; Universidad Tecnológica del Perú