A comparison of forecasting methods for hotel room occupancy

Abstract

There are a few types of forecasting categories that have been used such as hotel room occupancy forecast. Implementation of this forecasting category can be crucial because it leads to an efficient planning for, and decision making to all the hotel departments. Thus, this study aims to compare the best forecasting method for hotel room occupancy. Therefore, Seasonal Naive, Seasonal Holt Winter’s Method and ARIMA are going to be implemented in order to determine which forecasting method is most suitable to forecast hotel room occupancy by using secondary data from year 2012 until 2017. The selection of best method is based on three error measurements which are root mean square error (RMSE), mean absolute percentage error (MAPE) and mean absolute error (MAE). From the analysis conducted, the results show the best method to be implemented is the Seasonal Holt Winter’s Multiplicative method since it shows the lowest error for all three measurements. Furthermore, the forecast of future hotel room occupancy for year 2018 shows similar pattern as previous years. In comparing 2018 future occupancy with 2017 actual occupancy, there are some increment and decrement in hotel room occupancy for various months.There are a few types of forecasting categories that have been used such as hotel room occupancy forecast. Implementation of this forecasting category can be crucial because it leads to an efficient planning for, and decision making to all the hotel departments. Thus, this study aims to compare the best forecasting method for hotel room occupancy. Therefore, Seasonal Naive, Seasonal Holt Winter’s Method and ARIMA are going to be implemented in order to determine which forecasting method is most suitable to forecast hotel room occupancy by using secondary data from year 2012 until 2017. The selection of best method is based on three error measurements which are root mean square error (RMSE), mean absolute percentage error (MAPE) and mean absolute error (MAE). From the analysis conducted, the results show the best method to be implemented is the Seasonal Holt Winter’s Multiplicative method since it shows the lowest error for all three measurements. Furthermore, the forecast of future hotel room occupancy fo...