Jeju Island is one of South Korea’s most popular tourist destinations, attracting over 13 million visitors annually for various purposes. Depending on their individual needs, visitors choose different types of accommodations. Although online reviews are commonly used to obtain information about lodging facilities, analyzing such text-based data poses challenges due to its unstructured nature. Topic modeling has been widely used to analyze textual information and uncover latent themes, and it has also been applied in hotel review analysis. However, most previous studies have focused on analyzing the overall reviews from a specific website to identify general factors related to travel satisfaction. Moreover, the topic models used in these studies are limited in that they cannot incorporate external information beyond the text itself. This study aims to identify the characteristics of hotels in the Jeju region by utilizing hotel names provided alongside review texts. Review data for 12 hotels in Jeju Island were collected from the Agoda website. For analysis, we employed the neural topic model Scholar (Sparse Contextual Hidden and Observed Language Autoencoder), which is capable of incorporating external covariates. By including hotel names as external information in the Scholar model, we were able to identify not only latent topics but also hotel-specific characteristics. This analysis is expected to support travelers in selecting accommodations and contribute to improving the service quality of hotels.
Oh et al. (Sun,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: