PERANCANGAN SISTEM KLASIFIKASI SENTIMEN ULASAN PENGGUNA APLIKASI MYIM3 DENGAN METODE NAÏVE BAYES

MYIM3 application is a digital platform that provides various telecommunication services, such as credit purchase and data package purchase. As an internet network service provider company, with the increasing number of application users, users can provide reviews and ratings of products and services used by users. This requires a system that can automatically classify the sentiment of user reviews. This research aims to apply sentiment classification analysis to MYIM3 application user reviews sourced from the Google Play Store with a total of 9,989 sample data used with a time span from 2018-2024. The Naïve Bayes method is used to classify the sentiment of reviews into positive, neutral, or negative based on the star scores Given by users. Review data is preprocessed through cleaning, tokenization, filtering, and word weighting using TF-IDF. Modeling and evaluation of model performance were carried out using the Orange Data Mining platform with a division of training and test data of 80% and 20%. The results showed that the Naïve Bayes model was able to achieve 93.9% accuracy, 94.1% precision, 93.9% recall, and 93.9% F1-score. Factors such as data quality, data preprocessing, and dataset sharing techniques affect the performance of the model. Although there is still potential for improvement, the model performed well in classifying the sentiment of app user reviews.
Keywords: sentiment analysis, Naïve Bayes, classification, MYIM3, user reviews.