Table 2 Comparative training and testing outcomes for individual and hybrid models.
From: A hybrid SERWI ensemble model for crop yield prediction using an inverse RMSE weighting strategy
Model | Dataset | RMSE | MAE | R\(^2\) |
|---|---|---|---|---|
XGBoost | Train | 34.59 | 21.33 | 0.9973 |
Test | 45.87 | 38.34 | 0.9946 | |
Gaussian Process | Train | 104.80 | 81.22 | 0.9716 |
Test | 86.28 | 70.77 | 0.9876 | |
SVR | Train | 87.90 | 37.75 | 0.9825 |
Test | 38.69 | 24.34 | 0.9962 | |
Random Forest | Train | 85.80 | 59.43 | 0.9833 |
Test | 65.47 | 53.49 | 0.9890 | |
LSTM | Train | 70.35 | 35.00 | 0.9872 |
Test | 56.11 | 34.20 | 0.9947 | |
LSTM + SVR | Train | 70.74 | 39.89 | 0.9870 |
Test | 60.73 | 44.66 | 0.9938 | |
SVR + XGBoost | Train | 85.36 | 41.78 | 0.9811 |
Test | 67.77 | 45.20 | 0.9923 | |
XGBoost + LSTM | Train | 63.37 | 43.70 | 0.9896 |
Test | 74.96 | 56.18 | 0.9906 | |
LSTM + RF | Train | 78.24 | 66.50 | 0.9842 |
Test | 90.26 | 76.70 | 0.9864 | |
SERWI | Train | 64.56 | 34.97 | 0.9892 |
Test | 70.16 | 47.93 | 0.9918 |
