Enhancing Spam Detection on SMS performance using several Machine Learning Classification Models
Tarandeep Singh, Tushar Anupam Kumar, Prashant Giridhar Shambharkar · 2022 6th International Conference on Trends in Electronics and Informatics (ICOEI) · 2022
The Short Message Service (SMS) is swiftly turning into the most stable and secured mode of interaction because of its wide-ranging global coverage, reliability, dependability, and power efficiency. Because A2P messaging is more secure in comparison to P2P messaging, anybody can send a message, which leads to an attack. Spammers make use of this opportunity to spread harmful information, participate in disruptive behaviour, and harass others. In our work, we used machine learning classifiers for instance Multinomial-Naive-Bayes algorithm, Support-Vector-Machine models (SVM), Logistic-Regression, and Decision Trees to detect SMS spam from a dataset of nearly six thousand messages, taking into account combinations of Term-Frequency-Inverse-Document-Frequency (TF-IDF) and Count-Vectorization features to investigate the trade off between F1-score, accuracy -score, and computing time. Following that, we conducted a comparison study of the accuracy of various models for spam detection in order to determine the most accurate model that can be employed in this situation.