<![CDATA[The development of satellite communication technology has created opportunities for high-speed internet services across various regions, including remote areas. Starlink, which operates using low Earth orbit (LEO) satellites, and Telkom, which utilizes geostationary (GEO) satellites alongside terrestrial infrastructure, are two major providers in Indonesia. This study examines sentiment analysis of the service quality of Starlink and Telkom using a machine learning approach. Data were collected from the X (Twitter) platform and processed through cleaning, case folding, tokenization, slang word normalization, stopword removal, and stemming. Three classification algorithms were employed: Naïve Bayes, Decision Tree, and Random Forest. The results show that Decision Tree achieved the highest accuracy of 100%, followed by Random Forest at 96% and Naïve Bayes at 80%. Sentiment analysis revealed that Starlink was favored for its speed and connection stability, whereas Telkom was more appreciated for its affordability and service coverage. These findings provide an objective overview of user perceptions that can be utilized to improve service quality and strategic planning.]]>
Copyrights © 2025
