ABSTRACT
The accurate forecasting of hotel occupancy rates is crucial in most, if not all, departments of hotel operations. The natural advantages of a time series forecasting model are that the model is simple to apply and requires no more than a data series. The existing time series forecasting models either capture the information of the last few data in the data series or the entire data series is used for projecting future values. The time series forecasting models are unable to take advantage of the last trend in the data series, which always have a direct influence on the estimated values. This chapter proposes an improved extrapolative time series forecasting technique to compute future hotel occupancy rates. The performance of this new technique was tested with officially published room occupancy rates in Hong Kong. Forecasted room occupancy rates were compared with actual room occupancy rates in several accuracy performance dimensions.
