A COMPARISON OF THE FORECASTING FOR NUMBER OF FIRST-YEAR STUDENTS THE FACULTY OF EDUCATION VALAYA ALONGKORN RAJABHAT UNIVERSITY UNDER THE ROYAL PATRONAGE

Abstract

The purpose of this research was to study and compare forecasting methods for the number of first-year students at the Faculty of Education, Valaya Alongkorn Rajabhat University under the Royal Patronage, during the academic years 1997 to 2022, encompassing 26 values which were used and separated into 2 sets. The first set contained 20 values from the academic years 1997 to 2016 for the purpose of comparing and searching for forecasting models. The forecasting methods included Box-Jenkins, Holt’s exponential smoothing method, and the Damped trend exponential smoothing method. The second set consisted of 6 values from the academic years 2017 to 2022, used for comparison and determination of the most suitable forecasting method using the criteria of the lowest root mean squared error (RMSE) and mean absolute percentage error (MAPE). The results indicated that under the mean absolute percentage error (MAPE), the Damped trend exponential smoothing method was the most accurate. Meanwhile, for the root mean square error, the Box-Jenkins method with ARIMA (0,1,0) exhibited the best performance. Moreover, the forecast values from both methods were deemed reliable, and no statistically significant difference was observed between them.