ABSTRACT

The ubiquity of SMS text messaging has revolutionized mobile communication, lauded by the International Telecommunications Union (ITU) for its convenience. However, the proliferation of SMS spam poses a formidable challenge, necessitating robust detection mechanisms. In this study, we rigorously compare two Naïve Bayes variants—the Bernoulli Naïve Bayes Classification and the Multinomial Naïve Bayes Classification—using a substantial dataset of SMS messages. By evaluating their performance across diverse contexts, we contribute insights into their suitability for distinguishing spam from legitimate messages. Our findings enhance mobile communication security and show the unique characteristics of both models for SMS spam detection.