Table 2 Result of ANN in prediction of hybrid rice yield from their parent’s features.
From: Predicting hybrid rice performance using AIHIB model based on artificial intelligence
Hybrids | FS algorithm | Selected features | ANN structure | MSE (train) | MSE (vald) | MSE (test) | R2 (test) |
|---|---|---|---|---|---|---|---|
AHM × KHZ | GA | FLL, FLW, FLA, BI, FGN | 5–4–1 | 0.0348 | 0.0427 | 0.0275 | 0.9684 |
PSO | FLW, FLA, FGN, FLL | 4–8–1 | 0.0250 | 0.0673 | 0.0105 | 0.9694 | |
AHM × SPD | GA | PE, HE, PL, DFL, GY | 5–31–1 | 0.0981 | 0.1691 | 0.1262 | 0.9695 |
PSO | FLL, PL, DFL, PE, HE, GY | 6–31–1 | 0.0755 | 0.6932 | 0.1220 | 0.9694 | |
GHB × KHZ | GA | FGN, FLA, BI, PL | 4–6–1 | 0.0069 | 0.0038 | 0.0094 | 0.9688 |
PSO | FGN, FLL, FLA, BI | 4–9–1 | 0.0055 | 0.0093 | 0.0106 | 0.9682 | |
IR28 × GHB | GA | PE, GY, BI, HE, PL | 5–37–1 | 0.0086 | 0.0117 | 0.0088 | 0.9696 |
PSO | GY, PL, PE, BI, HE | 5–28–1 | 0.0054 | 0.0090 | 0.0090 | 0.9696 | |
IR28 × TAM | GA | FLA, PL, GY, HE, FLW | 5–40–1 | 0.0011 | 0.0126 | 0.0027 | 0.9697 |
PSO | PL, GY, HE, FLA, FLW | 5–41–1 | 0.0011 | 0.0046 | 0.0020 | 0.9698 | |
SHP × GHB | GA | GY, PBN, FLL, PL, HE | 5–33–1 | 0.0004 | 0.0020 | 0.0016 | 0.9695 |
PSO | GY, PBN, FLL, HE, PL | 5–42–1 | 0.0005 | 0.0040 | 0.0012 | 0.9694 | |
SHP × SPD | GA | FLL, GY, FLA, HE | 4–28–1 | 0.0055 | 0.0146 | 0.0056 | 0.9697 |
PSO | FLL, BI, HE, FLA, GY | 5–15–1 | 0.0060 | 0.0230 | 0.0080 | 0.9696 | |
TAM × KHZ | GA | DFL, FLW, FLA, FGN, BI | 5–30–1 | 0.0226 | 0.3924 | 0.0399 | 0.9658 |
PSO | FLW, BI, DFL, FLA | 4–19–1 | 0.0221 | 0.07027 | 0.0336 | 0.9678 | |
TAM × SHP | GA | FLA, PL, BI, GY | 4–34–1 | 0.0008 | 0.0011 | 0.0011 | 0.9697 |
PSO | FLL, BI, FLA, GY | 4–31–1 | 0.0009 | 0.0014 | 0.0013 | 0.9695 | |
General data | GA | PBN, FGN, HE, DFL, GY | 5–42–1 | 0.0486 | 0.0992 | 0.0738 | 0.9699 |
PSO | PBN, GY, DFL, FGN, HE | 5–32–1 | 0.0550 | 0.0518 | 0.0805 | 0.9698 |
