Table 2 The effect of each input parameter on the DeepESN model was evaluated by the value of γ(% change).
From: Application of the deep learning for the prediction of rainfall in Southern Taiwan
γ | Amount of change (%) | RMSE | |
|---|---|---|---|
Original | 0.507 | N/A | 1.51 |
Pressure | 0.355 | −30.05 | 1.8 |
Temperature | 0.489 | −3.63 | 1.56 |
Humidity | 0.463 | −8.70 | 1.57 |
Wind speed | 0.487 | −3.86 | 1.55 |
Wind direction | 0.499 | −1.60 | 1.52 |
Sea level | 0.501 | −1.16 | 1.52 |
Rainfall | 0.317 | −37.44 | 1.69 |
