Research Article
Electric Load Forecast Using Combined Models with HP Filter-SARIMA and ARMAX Optimized by Regression Analysis Algorithm
Table 8
The comparison of error among five models for training set.
| | Model | HP Filter-SARIMA-Revised ARMAX | SARIMA | ARIMA | Grey system | MLP neural network |
| | RMSE (root mean square error) | 21.47818 | 282.50409 | 433.36301 | 301.64787 | 224.33025 | | MAE (mean absolute error) | 18.4375 | 248.1875 | 396.375 | 261.9375 | 195.1875 | | MAPE (mean absolute percentage error) | 0.424 | 5.7531 | 9.2488 | 6.2738 | 4.5712 |
|
|
