<![CDATA[MyPertamina is PT Pertamina's digital application designed to support cashless and targeted distribution of subsidized fuel. Although it has been downloaded by millions of users, the app has received a variety of responses recorded in the form of reviews on the Google Play Store. This research was conducted to determine the tendency of user opinions through an automated sentiment analysis approach. Review data is collected by utilizing the Google Play Scraper library, then processed through preprocessing stages such as text normalization, character cleaning, tokenization, common word removal, and stemming. Positive and negative sentiment labels are assigned with the help of the IndoBERT pre-trained model. The next process includes conversion of text to numerical form with TF-IDF method and application of Multinomial Naïve Bayes algorithm for classification. The model is tested using confusion matrix and 10-Fold Cross Validation. The results showed that the majority of user reviews were negative (63.65%), and the classification model achieved 89.25% accuracy, 87.5% precision, 82% recall, and 84.69% F1-score. This shows that the approach used is effective in identifying public perceptions of MyPertamina application services.]]>
Copyrights © 2025
