Table 5.
Yield prediction performances of CNN-RNN(W), CNN-RNN(S), CNN-RNN(M), and Average model. Validation year is 2018. The unit of RMSE is bushels per acre.
| Response | Model | Training RMSE | Training Correlation Coefficient (%) | Validation RMSE | Validation Correlation Coefficient (%) |
|---|---|---|---|---|---|
| Corn | CNN-RNN(W) | 17.12 | 89.00 | 24.75 | 72.15 |
| CNN-RNN(S) | 19.13 | 84.70 | 24.41 | 72.82 | |
| CNN-RNN(M) | 26.77 | 69.49 | 33.03 | 32.99 | |
| Average | 37.51 | 0.0 | 35.56 | 0.0 | |
| Soybean | CNN-RNN(W) | 4.59 | 89.62 | 5.95 | 78.79 |
| CNN-RNN(S) | 5.60 | 83.25 | 5.93 | 81.23 | |
| CNN-RNN(M) | 7.86 | 59.86 | 8.67 | 48.78 | |
| Average | 10.28 | 0.0 | 10.10 | 0.0 |