Klasifikasi SMS Spam Berbahasa Indonesia Menggunakan Algoritma Multinomial Naïve Bayes

Abstract

Based on a report submitted by Truecaller Insights Report 2020, Indonesia placed sixth position with the most spam messages, one of the spam applications is SMS. Spam SMS contains unwanted or unsolicited messages, including advertisements, scams and so on. The existence of this spam message causes inconvenience from the user's side when receiving spam SMS, and some even become victims of crime after responding to the SMS. To minimize inconvenience and crime caused by spam messages, the purpose of this study is to filter SMS spam or SMS filtering by classifying SMS spam using the Multinomial Naïve Bayes algorithm by looking for the best combination of parameters to improve the performance of the model that is formed. The results of model testing get the highest precision value in the MNB and SVM models by 93%, the highest recall value in the SVM model at 94%, the highest f1-score value in the SVM model at 94%, the highest accuracy value in the SVM model at 95%, and the fastest test time on the MNB model is 2.66 ms