<![CDATA[SAMBARA is a digital service application developed by the West Java Provincial Government for motor vehicle tax payments, which has received a low rating of 2.4 from more than 39,000 reviews on the Google Play Store, making it a critical issue that undermines public trust in government digital services. This study analyzes 21,572 user reviews using a Natural Language Processing (NLP) approach with the Long Short-Term Memory (LSTM) algorithm through several preprocessing steps including cleaning, stopword removal, stemming, normalization, and Word2Vec numerical representation. The model was trained for 20 epochs with a batch size of 64 and evaluated using accuracy, precision, recall, and F1-score. The results show that negative reviews (46.8%) dominate over positive (45.9%) and neutral (7.3%) reviews, with the model achieving 81.1% accuracy and an F1-score of 0.81. Dominant negative words include “error” and “server”, while positive reviews highlight “tax” and “helpful”. These findings indicate that the main weakness of the application lies in technical aspects such as server stability, while its strength is in the core function of online tax payment. This study provides concrete recommendations for developers to improve system stability and offers methodological contributions for future research by adopting transformer-based models.]]>
Copyrights © 2025
