Skip to main content
Energy Informatics

Table 8 Comparison metrics with electric heating load

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

Model

MAE (kWh)

MAPE (%)

RMSE (kWh)

Adjusted R2

Electric heating and heat pumps with CSTEP variables

 FFN

2.8042

14.64

3.7258

0.8274

 RNN

2.7791

14.37

3.6699

0.8326

 LSTM

3.0864

16.07

4.1075

0.7902

 GRU

3.0004

15.52

3.9931

0.8018

Electric heating and heat pumps without CSTEP variables

 FFN

2.7671

14.33

3.6854

0.8311

 RNN

2.8852

15.36

3.7831

0.8221

 LSTM

3.0414

15.61

4.104

0.7906

 GRU

2.9284

15.47

3.8826

0.8126

Sampled aggregated load data (22 households)

 FFN

0.9375

11.01

1.3061

0.8273

 RNN

0.9212

10.99

1.2584

0.8397

 LSTM

1.0009

11.84

1.3922

0.8038

 GRU

1.0132

11.96

1.4121

0.7982

Back to article page