Classification of User Reviews of the JKN Mobile Application Using the Naive Bayes Algorithm

Abstract

The Mobile National Health Insurance (Mobile JKN) application is a digital service developed by BPJS Kesehatan to simplify access to administrative features and information related to the National Health Insurance (JKN) program. As its user base grows, the number of reviews on the Google Play Store continues to increase, providing important insights into user perceptions. However, these reviews have not been fully utilized to measure satisfaction with BPJS Kesehatan’s digital services. This study aims to classify user sentiment toward Mobile JKN automatically by analyzing reviews obtained through web scraping from the Google Play Store. The reviews were processed through several text preprocessing steps, including case folding, tokenization, stopword removal, and stemming, then transformed into numerical features using the TF-IDF method. The sentiment classification process uses the Naive Bayes Classifier, which is effective for large-scale text processing. Evaluation results show strong performance, achieving an accuracy of 92.26%, precision of 94.69%, recall of 92.81%, and an F1-score of 93.74%. Most reviews express positive sentiment and highlight the application's usefulness, while negative reviews generally point to technical problems such as login errors and system instability