A Study on Regional Hotel Demand using Machine Learning Techniques

The purpose of this study is to find regional hotel demand in the midst of the recovery in tourist numbers following COVID-19 and to provide suggestions. In addition, it employs machine learning techniques, which has recently become popular, to draw regional hotel demand in the country and thereby make suggestions. To predict regional hotel room demand, it used machine learning models in which exchange rates, real GDP, economically active population, trade volume, and the number of foreign arrivals were set as input variables. As a result, the national supply of hotel rooms was 28,499 thousand, and the support vector machine model predicated an oversupply until 2030, whereas XGBoost model forecasted a shortage of 761 thousand rooms in 2029 and 2,027 thousand rooms in 2030. According to the support vector machine based supply analysis by city and province, the machine learning model predicted a shortage in Ulsan and Chungcheongnam-do after 2025, in Gangwon and Jeju-do after 2026, and in Seoul and Daegu after 2029. The XGBoost based supply analysis revealed that it predicted a shortage in Ulsan and Chungcheongnam-do after 2025, in Seoul after 2027, in Daejeon after 2028, in Busan, Incheon and Jeju-do after 2029.