<![CDATA[The rapid growth of online platforms has enabled users to share their experiences about various products and services, including hotels. Hotel reviews are crucial in understanding customer perceptions and preferences in the tourism sector. Tiket.com, a web and mobile-based online travel agent, allows users to book hotels and submit reviews, which can be positive, negative, or neutral. These reviews provide valuable insights into the strengths and weaknesses of hotel services and can serve as evaluation material for improvements. This study extracts meaningful information from user reviews through an aspect-based sentiment analysis approach. It categorizes sentiments into specific aspects such as price, cleanliness, service, location, and facilities, ensuring the feedback is more structured and actionable. The research utilizes a Gated Recurrent Unit (GRU) model combined with fastText word embedding to analyze sentiment. A dataset of 6512 hotel reviews was collected through web scraping. The resulting model achieved an accuracy of 91%, evaluated using a confusion matrix. The approach enhances understanding of customer satisfaction by presenting sentiments based on targeted service aspects, making the analysis more concise and relevant for hotel management.]]>
Copyrights © 2025
