Table 5.
Method ††† | Pear Cultivars | Training | Test | Reference | |
---|---|---|---|---|---|
R2 | R2 | RMSE (g kg−1) | |||
Linear regression | Rocha, Huanghua | 0.87 | 0.54 to 0.99 | No detail data | Neto et al., 2011; Yang et al., 2011 |
PLSR | Cuiguan, Huangguan | 0.90 | 0.72 | 2.95 | Wang et al., 2014 |
Vegetation index | Kotobuki shinsui, Red-blush | 0.46–0.67 | 0.41–0.51 | 3.0–3.35 | Wang et al., 2017; Perry et al., 2018 |
PLSR | Kotobuki shinsui | 0.86 | 0.85 | 1.50 | Wang et al., 2017 |
NN | Kotobuki shinsui | 0.89 | 0.67 | 1.70 | Wang et al., 2017 |
Vegetation index | Mixed cultivars | 0.45 | 0.42 | 4.55 | This paper |
PLSR | Mixed cultivars | 0.85 | 0.76 | 3.46 | This paper |
NN | Mixed cultivars | 0.95 | 0.85 | 1.66 | This paper |
AdaBoost.RT-BP | Mixed cultivars | 0.97 | 0.92 | 1.29 | This paper |
††† DSI, PLSR, and NN represent difference spectral index, partial least squares regression, and neural networks, respectively.