Table 1 Confusion matrices and metrics of the interactive classification of soybean seeds according to their visual appearance.
From: Interactive machine learning for soybean seed and seedling quality classification
Classa | High-quality seed | Kneaded seed | Purple stained seed | Broken seed | Seed coat tear | Moisture damaged seed | Green-ish seed |
|---|---|---|---|---|---|---|---|
n = 630 | |||||||
High-quality seed | 70 | 1 | 3 | 0 | 2 | 27 | 0 |
Kneaded seed | 0 | 76 | 17 | 0 | 0 | 0 | 6 |
Purple stained seed | 1 | 2 | 64 | 0 | 0 | 0 | 2 |
Broken seed | 0 | 0 | 0 | 87 | 0 | 1 | 0 |
Seed coat tear | 0 | 3 | 3 | 3 | 88 | 1 | 2 |
Moisture damaged seed | 19 | 0 | 1 | 0 | 0 | 61 | 3 |
Greenish seed | 0 | 8 | 2 | 0 | 0 | 0 | 77 |
Accuracy | 0.92 | 0.95 | 0.94 | 0.99 | 0.98 | 0.92 | 0.96 |
Kappa | 0.68 | 0.78 | 0.77 | 0.97 | 0.91 | 0.65 | 0.85 |
Precision | 0.68 | 0.77 | 0.93 | 0.99 | 0.88 | 0.73 | 0.89 |
Sensitivity | 0.78 | 0.84 | 0.71 | 0.97 | 0.98 | 0.68 | 0.86 |
Specificity | 0.94 | 0.96 | 0.99 | 1 | 0.98 | 0.96 | 0.98 |
