Predicting Hotel Economic Trends Using Machine Learning and Time Series Analysis Models

Abstract

(Purpose) The purpose of this study is to construct a time-series forecast model for the hotel service production index using the ARIMA model to predict the hotel industry's economic trend, and to analyze its long-term growth and short-term volatility. This study aims to help hotel executives make strategic decisions to effectively respond to a rapidly changing market environment. (Design/methodology/approach) This study constructed the ARIMA model using 16 years of time-series data provided by KOSIS and the Korea Hotel Association. Through this, the service production index forecast of the hotel industry was carried out, and various statistical indicators were used to evaluate the accuracy of the forecast model. In addition, the prediction results were revalidated using various test datasets to improve the model's performance. (Findings) The ARIMA model has shown a high degree of suitability as a tool that can accurately predict not only the long-term growth trends of the hotel industry, but also short-term volatility. In particular, it was confirmed that the model can respond effectively in the event of unexpected economic shocks such as COVID-19. It provides a strategic foundation for hotel management to respond nimbly to market changes.. (Research implications or Originality) In order to overcome the limitations of the time-series prediction model, this study developed a prediction model optimized for the hotel industry by comparing and analyzing machine learning techniques. In addition, customized forecasting models that can be applied to other service industries will be presented to provide basic data for related research.