Research Article
A Comparison of Hour-Ahead Solar Irradiance Forecasting Models Based on LSTM Network
Table 9
Performance comparison between the proposed LSTM-MLP model and the general machine learning method.
| Model | RMSE (W/m2) | nRMSE (%) | MAE | MBE | R |
| BPNN | 76.9272 | 39.9215 | 33.2872 | −4.7798 | 0.9601 | RNN | 77.8444 | 40.4115 | 40.0704 | −16.2864 | 0.9602 | Random forest | 70.3808 | 36.5369 | 30.5618 | −0.9593 | 0.9659 | SVM | 77.0384 | 39.9931 | 50.9564 | −1.6776 | 0.9593 | General LSTM | 71.8434 | 37.2962 | 30.6385 | −1.9523 | 0.9649 | The proposed LSTM-MLP | 62.1618 | 32.2702 | 26.6538 | −0.4547 | 0.9737 |
|
|