Leveraging large language models for daily tourist demand forecasting

Large Language Models have attracted the attention of tourism researchers, and many discussions have been published in leading tourism journals. However, little research has been conducted on how to use Large Language Models in tourism research. In this paper, we propose a new tourist arrival forecasting model based on the Large Language Model. The Large Language Model is used to extract and produce the satisfaction scores from the review comments by the tourists. The generated satisfaction scores are incorporated into the tourist arrival forecasting model to take full advantage of the extracted information from the review comments. We have applied the Large Language Model based forecasting model to predict the tourist arrival in Macao. Experiment results show that the proposed model has produced forecasts with improved forecasting accuracy.