Table 1 Results of parameter sensitivity experiments using MobileNetV3-Small on the first fold of the dataset
From: Phenotypic feature-based identification of tea geographical origin using lightweight deep learning
Epochs | Learning rate | OA | AA | \(\kappa\) | MCC | Precision | F1 |
|---|---|---|---|---|---|---|---|
100 | 0.001 | 0.82 | 0.85 | 0.79 | 0.79 | 0.85 | 0.82 |
100 | 0.01 | 0.63 | 0.66 | 0.56 | 0.57 | 0.65 | 0.63 |
200 | 0.001 | 0.56 | 0.54 | 0.47 | 0.48 | 0.54 | 0.53 |
200 | 0.01 | 0.78 | 0.79 | 0.74 | 0.74 | 0.79 | 0.77 |
300 | 0.001 | 0.69 | 0.73 | 0.63 | 0.63 | 0.70 | 0.68 |
300 | 0.01 | 0.76 | 0.81 | 0.71 | 0.72 | 0.78 | 0.76 |
