Short Message Service (SMS) Spam Filtering using Machine Learning in Bahasa Indonesia
Agustinus Theodorus, Tio Kristian Prasetyo, Reynaldi Hartono, Derwin Suhartono
Abstract
Short Message Service (SMS) is an essential communication tool in Indonesian society. Companies use SMS as a promotion tool but unfortunately some individuals use SMS to send spam messages. A smartphone user in Indonesia has had an experience with these spam and promotional messages. This study presents a model to classify spam, promotion and ham messages based on Indonesian text messages. The model was trained with 4,125 text messages, tested with 1,260 text messages. A 10-fold cross validation method was used to evaluate the classifiers and the results show that Random Forest (94.62%), Multinomial Logistic Regression (94.57%), Support Vector Machine (94.38%), and XGBoost (94.52%) are among the best models to be used for a multiclass SMS classification.