Abstract— Short message service (SMS) is one of the important
communication media to support the speed of using mobile
phones by users. The SMS classification hybrid system is used to
detect SMS that are considered junk and correct. In this
research, what is needed is to collect SMS datasets, feature
selection, preprocessing, vector creation, filtering and updating
the system. Two types of SMS classification on mobile phones are
currently listed as blacklisted (rejected) and whitelisted
(accepted). This study uses several algorithms such as support
vector machine, Naïve Bayes classifier, Random Forest and
Bagging Classifier, which aims to produce an algorithm that has
the highest performance score and is accurate in filtering
incoming SMS. In this study, it was found that the Bagging
classifier algorithm got the highest performance score from other
algorithms which can be used as a means to filter incoming SMS
into the user's inbox and the Bagging classifier algorithm so as to
provide accurate filtration results to filter incoming SMS.