Research Article
Predicting Sea Surface Temperatures in the North Indian Ocean with Nonlinear Autoregressive Neural Networks
Table 2
NAR network testing at site AS.
| Prediction horizon in months | Testing phase CC | MAE in °C | MSE in °C2 | NSE in % |
| | 0.99 | 0.15 | 0.04 | 96.90 | | 0.96 | 0.28 | 0.13 | 90.70 | | 0.96 | 0.27 | 0.12 | 91.20 | | 0.95 | 0.29 | 0.15 | 89.50 | | 0.95 | 0.28 | 0.14 | 90.10 | | 0.96 | 0.28 | 0.14 | 90.20 | | 0.96 | 0.28 | 0.14 | 90.30 | | 0.96 | 0.30 | 0.14 | 89.90 | | 0.96 | 0.28 | 0.13 | 90.10 | | 0.95 | 0.30 | 0.15 | 89.20 | | 0.95 | 0.30 | 0.15 | 89.30 | | 0.96 | 0.28 | 0.12 | 91.00 |
|
|
Note: : SST at th time (month) ahead than the present time (month) “”.
|