Skip to main content
. 2021 Nov 6;14:4641–4655. doi: 10.2147/IDR.S337473

Table 3.

Measures of Predicted Performance for BSTS Methods and ARIMA Methods

Models Testing Horizons Testing Horizons
MAD MAPE RMSE MER RMSPE MAD MAPE RMSE MER RMSPE
12-Step ahead prediction 60-Step ahead prediction
ARIMA 6403.506 7.671 7212.042 0.074 0.088 11,711.524 12.960 13,859.798 0.127 0.154
BSTS 6024.076 7.368 6848.671 0.070 0.086 4846.300 5.400 6353.691 0.052 0.072
Reduced percentage (%) Reduced percentage (%)
BSTS vs ARIMA 5.925 3.950 5.038 5.405 2.273 58.619 58.333 54.157 59.055 53.247
36-Step ahead prediction 96-Step ahead prediction
ARIMA 5692.866 6.234 6838.697 0.063 0.075 12,958.062 13.848 15,136.311 0.132 0.166
BSTS 4733.059 5.182 6250.588 0.052 0.068 7077.838 7.297 8924.741 0.072 0.091
Reduced percentage (%)
BSTS vs ARIMA 16.860 16.875 8.600 17.460 9.333 45.379 47.306 41.038 45.455 45.181

Abbreviations: BSTS, Bayesian structural time series method; ARIMA, autoregressive integrated moving average method; MAD, mean absolute deviation; MAPE, mean absolute percentage error; RMSE, root-mean-square error; MER, mean error rate; RMSPE, root-mean-square percentage error.