Table 4.

Out-of-sample forecasting accuracy for residents

6-steps ahead12-steps ahead
ModelMAEMCSRMSEMCSMAEMCSRMSEMCS
Period 1        
SSARIMA36.69310.039.15260.046.93850.050.19910.0
MSARIMA40.92710.043.11910.050.86210.053.31580.0
ANN34.43740.038.17270.045.57480.049.02980.0
LSTM25.11751.030.67451.029.03281.033.70291.0
RNN34.96850.038.95380.045.60260.049.91260.0
SVM38.55270.042.55190.049.45690.053.16820.0
Naïve 142.75890.045.36500.050.84610.054.07160.0
Period 2        
SSARIMA16.53030.3118.26010.2517.03490.018.82420.0
MSARIMA19.51200.020.54290.019.49280.022.70980.0
ANN27.06160.028.99980.029.62350.032.68260.0
LSTM16.43831.019.60391.013.89281.015.32191.0
RNN32.28640.034.58930.029.71500.032.75710.0
SVM22.86430.024.91710.022.97370.024.97300.0
Naïve 132.21210.033.16620.030.92310.034.85000.0
Period 3        
SSARIMA26.77980.031.66630.021.87930.026.05920.0
MSARIMA24.18950.028.27250.022.74250.026.32760.0
ANN28.41340.033.82110.025.13110.028.16440.0
LSTM22.75171.024.93441.013.40121.014.99471.0
RNN23.61510.2525.31580.3118.40090.020.99790.0
SVM30.15440.035.12580.027.11830.029.98090.0
Naïve 131.75450.036.11850.030.38460.032.52210.0

Note:

The italic numbers indicate the lowest error rate (MAE and RMSE)

or Create an Account

Close Modal
Close Modal