A Collective Approach to SMS Spam Detection: Harnessing the Power of Multimodal Features, Machine Learning, and Advanced Classifiers
B. Sandhya, Darshan R.M, L. Anushree · 2023
In the current of mobile communication, Short Message Service (SMS) remains a dominant means of interaction despite the proliferation of messaging apps. Unfortunately, this increase in SMS traffic has given rise to an increase in SMS spam. This unwelcome phenomenon has become a conduit for advertising and fraudulent activities, causing significant inconvenience and harm to users. Thus, achieving a better way of SMS spam filtering is a formidable challenge. In this research, a machine learning based approach to spam SMS detection is used. This approach uses a variety of machine learning algorithms, including Logistic Regression, Support Vector Machines (SVMs), Multinomial Naive Bayes (MNB), Voting Classifiers, Stacking Classifiers and many other ML models. The best model based on its performance, measured by accuracy and precision is found out to be Multinomial Naïve Bayes, this model is chosen and further used for spam and ham classification. This model resulted in accuracy of 97% and precision of 100%. This shows that multinomial naïve bayes is the most suitable model for this classification task.