WSEAS Transactions on Computer Research
Print ISSN: 1991-8755, E-ISSN: 2415-1521
Volume 14, 2026
Time Series Models and Machine Learning for Predicting Tourism Demand in Emerging Destinations
Authors: , , ,
Search Articles
Abstract: In this study, we used time series models and machine learning techniques to predict tourism demand for emerging destinations. We collected data on visitors, weather, room rates, things to do, online activities, and relationships from the beginning of 2017 to 2023. For this study, we used ARIMA, SARIMA, and LSTM models. We found that LSTM provides the most accurate predictions with the lowest mean absolute error (MAE) and root mean square error (RMSE). In situations with sudden and frequent changes, LSTM is the superior model. ARIMA is acceptable but simplistic. SARIMA took seasonal data into account but showed the highest error and failed to fit the data. The study showed that AI tools, including LSTM, can play a critical role in tourism investment, especially where conventional models have limited applications. The study recommends their use in already available tourism models.
Keywords:
Tourism demand, Emerging destinations, Time series, Machine learning, ARIMA models, SARIMA models, LSTM networks
Pages: 471-478
DOI: 10.37394/232018.2026.14.41