Table 2 Performance of PV power forecasting models on testing set.

From: A novel hybrid model integrating CEEMDAN decomposition, dispersion entropy and LSTM for photovoltaic power forecasting and anomaly detection

Model

RMSE

MAE

SSE

nRMSE

nMAE

R²

GRU

0.554

0.481

5388.6

7.72%

6.71%

0.856

LSTM

0.509

0.390

4558.0

7.10%

5.44%

0.878

Transformer

0.361

0.212

2282.6

5.00%

3.00%

0.938

DLinear

0.358

0.200

2246.3

5.00%

2.80%

0.939

CEEMDAN-LSTM

0.337

0.153

1994.4

4.70%

2.14%

0.946

CEEMDAN-DispEn-LSTM

0.326

0.112

1868.9

4.54%

1.67%

0.950

  1. The bold values are the optimal values among all methods.
Back to article page