Smishing: A SMS Phishing Detection Using Various Machine Learning Algorithms

Amid the pandemic, there has been a steep rise in cybercrimes against individuals and corporations, making implementing security measures even more imperative. This paper proposes a machine learning-based approach for detecting phishing SMS threats using datasets of manually made SPAM and HAM texts. Additionally, pre-existing link datasets are used for training and testing spam and ham links. Furthermore, cloud-hosted application is developed for proof of concept, capable of detecting malicious URLs and SMS. VirusTotal API is utilized and integrated with the application for detecting harmful URLs using existing datasets. The datasets are evaluated using random forest (RF), long short-term memory (LSTM), logistic regression (LR), and support vector machine (SVM) algorithms conducted by considering precision, recall, and F1-score, ensuring efficiency in distinguishing between legit and spam. This paper enhances SMS phishing protection using machine learning advancements, demonstrating robust defense against phishing attempts, suggesting widespread integration into mobile security frameworks.