Sentiment Analysis of User Reviews for Health Technology Applications (Halodoc, Alodokter, and SatuSehat) Using the Multinomial Naive Bayes Algorithm
International Journal of Management Science and Information Technology · 17.10.2026 · 6 (2) · 2872-2882
DOI: 10.35870/ijmsit.v6i2.8850 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Management Science and Information Technology, online erschienen am 17.10.2026. Verfasst von Adhi Bahtiar und 2 weiteren Autor:innen. Die Arbeit nennt 12 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
The growth of digital health services in Indonesia has driven increased usage of health applications such as Halodoc, Alodokter, and SatuSehat. These applications generate thousands of user reviews on the Google Play Store annually. As these reviews consist of unstructured text, manual sentiment identification is time-consuming and labor-intensive. This study aims to perform automated sentiment analysis on user reviews for these three applications using the Multinomial Naive Bayes algorithm combined with Term Frequency-Inverse Document Frequency (TF-IDF) weighting. The research employs the Cross-Industry Standard Process for Data Mining (CRISP-DM) framework, encompassing data collection via scraping; text preprocessing (case folding, cleaning, tokenizing, stopword removal, and stemming); initial labeling using the InSet lexicon; TF-IDF weighting with a combination of unigrams and bigrams; Multinomial Naive Bayes model training; and result presentation via a Streamlit-based interactive dashboard. The final dataset comprises 10,739 reviews, split into 80% training data and 20% testing data. Test results indicate that the model achieved an accuracy of 48.3%, precision of 91.6%, recall of 48.3%, and an F1-score of 59.8%. The best performance was observed in the positive class (precision of 0.98), whereas the neutral class showed poor performance due to data imbalance. Analysis results reveal that SatuSehat had the highest proportion of negative sentiment (70.1%), while Halodoc had the highest proportion of positive sentiment (88.6%). This study provides insights into user perceptions of the three applications and offers valuable information for developers to improve their services.
Zitieren
APA
Bahtiar, A., Darmanto, E., & Arifin, M. (2026). Sentiment Analysis of User Reviews for Health Technology Applications (Halodoc, Alodokter, and SatuSehat) Using the Multinomial Naive Bayes Algorithm. International Journal of Management Science and Information Technology, 6(2), 2872-2882. https://doi.org/10.35870/ijmsit.v6i2.8850
Vancouver
Bahtiar A, Darmanto E, Arifin M. Sentiment Analysis of User Reviews for Health Technology Applications (Halodoc, Alodokter, and SatuSehat) Using the Multinomial Naive Bayes Algorithm. International Journal of Management Science and Information Technology. 2026;6(2):2872-2882. doi:10.35870/ijmsit.v6i2.8850
BibTeX
@article{Bahtiar2026,
author = {Bahtiar, Adhi and Darmanto, Eko and Arifin, Muhammad},
title = {Sentiment Analysis of User Reviews for Health Technology Applications (Halodoc, Alodokter, and SatuSehat) Using the Multinomial Naive Bayes Algorithm},
journal = {International Journal of Management Science and Information Technology},
year = {2026},
volume = {6},
number = {2},
pages = {2872--2882},
doi = {10.35870/ijmsit.v6i2.8850}
}
EUVIMED