Topic Modeling and Text Clustering of Foreign Tourists' Reviews on Khao Yai National Park Using Unsupervised Learning

Abstract

This research on topic modeling combined with text clustering of international tourists’ reviews on Khao Yai National Park utilizing unsupervised learning techniques collected a total of 3,491 reviews from various online platforms between January 1, 2023, and December 31, 2024. The data underwent a cleaning process before being transformed into numerical vectors using the Term Frequency-Inverse Document Frequency (TF-IDF) technique. Topic modeling was then conducted using the Latent Dirichlet Allocation (LDA) method. The analysis identified five main topics: 1) tourist attractions and the natural environment, 2) service quality and available facilities, 3) costs and value for money, 4) transportation and accessibility, and 5) personal experiences and enjoyment. These results correspond with the findings from K-means clustering of negative reviews, which were categorized into three major groups: 1) entrance fees and pricing structures, 2) accessibility and service satisfaction, and 3) facilities and tourism activities. These factors were found to significantly influence tourist satisfaction. Addressing the identified shortcomings can support the development of appropriate strategies to meet the diverse needs of tourists. Furthermore, the application of big data in systematic policy development is instrumental in shaping responsive management approaches that effectively tackle emerging issues and contribute to the long-term sustainability of Thailand’s tourism industry.