SMS SPAM FILTER using multinomial naïve bayes approach
Vijay Kumar Sharma, Vedant Garg, Sparsh Goel, Rohan Kapoor, Shivansh Singh · 2025
Unsolicited electronic messages sent via text, commonly referred to as SMS spam, involve the practice of sending unwanted texts for various purposes, often for commercial gain or engaging in illicit activities. Despite the existence of numerous anti-spam measures, SMS spam remains a prevalent issue on mobile devices due to its cost-effectiveness and the significant impact of spam campaigns. Various solutions, such as white-listing, grey-listing, blacklisting etc., are employed to categorize incoming messages [9]. However, a definitive solution is yet to be established. One potential reason for this challenge is the resilience of spammers; they adapt to and overcome spam filtering methods when compromised. The goal of this research is to enhance the detection of SMS spam using the Multinomial Naïve Bayes approach, coupled with text sanitation and TF-IDF. The results from the proposed model indicate an improvement in accuracy and precision scores as compared to using Multinomial Naïve Bayes alone.