Network | HerK |
Kp
|
---|
| RMSE | CC | RMSE | CC |
---|
4:4:1 | 10.8416 | 0.71 | 9.4223 | 0.87 |
4:5:1 | 10.1515 | 0.76 | 9.3945 | 0.87 |
4:6:1 | 10.3427 | 0.75 | 9.1942 | 0.88 |
4:7:1 | 10.1915 | 0.76 | 9.4011 | 0.88 |
4:8:1 | 10.3581 | 0.75 | 9.3961 | 0.87 |
- This table shows different NN configurations that were investigated with the corresponding prediction performance (on both K and Kp indices), evaluated by calculating the RMSE (RMSE values in this table are multiplied by 10) and the CC over the whole validation data set. The three numbers (i.e. 4:5:1) in the first column indicate the number of input, hidden and outputs layers, respectively.