Forecasting Air Passengers in Phuket, Thailand: An Application of Machine Learning Algorithms

The COVID-19 pandemic occurred in 2020 had a significant impact on the global economy, resulting in a significant decrease in air passengers. As a result, time series trends experienced significant fluctuations that were difficult to predict. The objective of this research was to forecast the number of air passengers in Phuket, Thailand, from September to December 2022. Additional features, such as exchange rates and time-based properties, were included to enhance the accuracy of our predictions. We then compared the performance of statistical models and machine learning models using evaluation metrics such as mean absolute error (MAE), root mean squared error (RMSE), and mean absolute percentage error (MAPE). Our findings revealed that the SARIMAX model outperformed other models for air passenger forecasting. Nonetheless, machine learning models might be more appropriate for forecasting new data due to their stability with additional training data. This research emphasized the importance of selecting the appropriate model based on the available data and the particular needs of the forecasting problem.