Table 2.

Mean Absolute Percentage Error (MAPE) [Root Mean Square Error (RMSE)] values of the best-fitting forecasting models and the fixed average census approach for different census periods during January – December 2013. MAPE values are measured as percentages. RMSE values are measured in number of patients.

Model	January 14 – 20, 2013 (High Census)	April 1 – 7, 2013 (Low Census)	May 3 – 9, 2013 (Average Census)	Average MAPE [RMSE] of All Predictions
Fixed average census approach	0.192 [12.294]	0.180 [7.681]	0.038 [2.035]	0.119 [6.272]
ARIMA(1,0,0)	0.056 [4.016]	0.080 [3.636]	0.059 [3.662]	0.074 [3.775]
ARIMA(1,0,0)x(1,1,2)7	0.042 [3.420]	0.098 [4.288]	0.062 [3.618]	0.073 [3.686]
ARIMA(2,1,4)x(1,1,2)14	0.039 [3.211]	0.078 [3.462]	0.067 [3.912]	0.081 [4.090]
Linear Regression	0.075 [5.149]	0.059 [2.755]	0.063 [3.331]	0.071 [3.622]