<![CDATA[This research examines user responses concerning public service applications from the Indonesian government. The information collected from the Google Play Store between January 1, 2025, and October 31, 2025, final dataset consists of 7,000 reviews that were labeled manually. The primary goal is to apply Aspect-Based Sentiment Analysis (ABSA) via E-S-QUAL to assess sentiment across four dimensions (efficiency, fulfillment, system availability, and privacy). In the classification process, IndoBERT Fine-Tuned model's efficacy was compared to traditional models (SVM and Naïve Bayes, showing superiority by reaching an overall accuracy of 91%, positioning it as a benchmark for all sentiment analysis. The model then directed the BERTopic analysis to pinpoint and explain the specific issues raised by users. The general results indicate that user sentiment towards the applications is predominantly Negative (62.71%). Through the application of topic modeling and E-S-QUAL mapping, it was found that System Availability (58.24% Negative) and Fulfillment (78,32% Negative) are the key contributors to dissatisfaction, arising from particular grievances such as system outages, service disruptions, and Privacy aspect (91.67% Negative) that concerns users. Conversely, the Efficiency aspect emerges as a crucial service strength, garnering a dominant positive reaction (58.96%). Therefore, these findings clearly suggest that the government must prioritize immediate action to improve system stability and ensure consistent delivery of services, thereby realizing the full potential of its digital transformation efforts. ]]>
Copyrights © 2026
