Short Message Service Filtering with Natural Language Processing in Indonesian Language

Vincentius Gabriel Tandra, Yowen Yowen; Ravel Tanjaya; William Lucianto Santoso; Nunung Nurul Qomariyah

As the amount of spam on messaging platforms such as emails have increased, the same has happened within Short Message Service (SMS) services as well. Within this study, Natural Language Processing was used on SMS in Indonesian Language (Bahasa), to create an Artificial Intelligence (AI) model capable of distinguishing between spam and other types of messages that are not spam. Within this study, we compared the performance of the Multinomial Naive Bayes Classifier and the Bi-Directional LSTM algorithm. We demonstrated this using code written in Python and the TensorFlow and Scikit libraries to generate reports, graphs and an application to test the performance of the models. Our results reveal that these methods are effective in filtering Bahasa Indonesia spam within SMS inboxes. In addition, we also published the SMS dataset in Bahasa with this paper.