Investigating Evasive Techniques in SMS Spam Filtering A Comparative Analysis of Machine Learning Models

This study addresses the challenges of SMS spam detection by analysing and comparing the effectiveness of various machine learning models on a Kaggle-sourced dataset containing both spam and non-spam messages. We employed Decision Tree, Random Forest, Roberta, and Distil Roberta algorithms to discern patterns and distinguish between legitimate and unsolicited messages. Each model was evaluated based on its accuracy, precision, and recall in identifying spam. The data consisted of labelled messages, with 'ham' indicating non-spam and 'spam' for unsolicited content. The objective was to identify which models performed best in terms of detection rates and resilience against evasion techniques used by spammers. Our analysis aimed to provide insights into the adaptability of advanced machine learning models in combating the evolving tactics of spam in SMS communications.

Keywords: Decision Tree, Random Forest, Roberta, and Distil Roberta.