Skip to main content

Table 6 One-hour forecast metrics

From: Evaluation of neural networks for residential load forecasting and the impact of systematic feature identification

Model

MAE (kWh)

MAPE (%)

RMSE (kWh)

Adjusted R2

With CSTEP variables

 FFN

3.9064

4.65

5.2668

0.9681

 RNN

4.1338

5.14

5.4095

0.9663

 LSTM

3.9887

4.72

5.4451

0.9659

 GRU

4.0633

4.91

5.3968

0.9665

Without CSTEP variables

 FFN

5.1670

6.43

6.6406

0.9494

 RNN

4.3655

5.49

5.6194

0.9637

 LSTM

3.9674

4.71

5.3706

0.9669

 GRU

4.0991

4.94

5.4944

0.9653