| dc.description.abstract | The rapid increase in the usage of online platforms has allowed users to share their experiences and opinions about various products and services, including hotels. The hotel industry is one of the continuously growing sectors in the tourism industry. In this study, the analysis of hotel reviews becomes crucial to understand customers' perceptions and preferences towards the provided services. Tiket.com is one of the companies operating in the tourism sector, serving as an online travel agent with a web and mobile-based application. Users can use this application to book hotels and provide reviews, whether they are neutral, negative, or positive. From the reviews given by these users, the hotel can find out the advantages and disadvantages of their hotel, so that these reviews can be used as an evaluation. This research aims to extract information obtained from user sentiment by using an aspect-based sentiment analysis approach to categorize each aspect, such as price, cleanliness, service, location, and facilities mentioned in the hotel reviews. This method ensures that the obtained information is concise, clear, and provides a deeper understanding of customer satisfaction. The Gated Recurrent Unit (GRU) and word embedding fastText are employed in this research. The dataset used in this study consists of 6512 data, which were collected by scraping from one of the hotel booking applications. The built model achieved an accuracy rate of 91% based on the five predetermined aspects, with the application of the confusion matrix evaluation method. | en_US |
