Table 4

Comparison of in-sample fitting and out-of-sample predicting performances among the best-performing approaches chosen

ModelsSimulating powerPredictive power
MAEMERMAPERMSEMAEMERMAPERMSE
 SARIMA5636.3030.0620.0678781.1865726.2620.0610.0607104.34
 SARIMA-GRNN4437.9580.0490.0546939.0785415.9850.0580.0596155.964
 SARIMA-NARNN3283.2740.0350.0435265.825259.5560.0560.0556418.445
 SARIMA-NARNNX2878.4840.0310.0384468.5783563.1790.0380.0384917.829
Percentage reductions (%)
 D versus A48.93050.00043.28449.11237.77537.70536.66730.777
 D versus B27.66829.03223.88128.13432.35632.78735.00017.428
 D versus C7.1826.4527.4639.07929.62529.50828.33321.123
  • GRNN, generalised regression neural network; MAPE, mean absolute percentage error; MER, mean error rate; MAE, mean absolute error; NARNN, non-linear autoregressive neural network; NARNNX, non-linear autoregressive neural network with exogenous input; RMSE, root mean square error; SARIMA, seasonal autoregressive integrated moving average.