Table 1.
PCA-LDA classification model on the NIRS data of pure and adulterated mixtures after three-fold cross-validation (N = 268).
| Validation Accuracy % | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Robusta-to-Arabica Ratio | |||||||||||
| 0% | 0.5% | 1% | 2% | 3% | 5% | 10% | 20% | 35% | 100% | ||
| Robusta-to-Arabica ratio | 0% | 66.74 | 8.03 | 0 | 7.44 | 0 | 11.11 | 11.11 | 0 | 0 | 0 |
| 0.5% | 14.79 | 51.92 | 7.45 | 7.44 | 14.78 | 7.44 | 7.44 | 0 | 3.67 | 0 | |
| 1% | 0 | 3.96 | 70.41 | 0 | 7.44 | 0 | 0 | 11.11 | 0 | 0 | |
| 2% | 0 | 8.03 | 0 | 81.44 | 7.44 | 0 | 0 | 0 | 0 | 0 | |
| 3% | 3.67 | 20.02 | 3.67 | 0 | 55.56 | 7.44 | 3.67 | 0 | 0 | 0 | |
| 5% | 3.67 | 0 | 3.67 | 3.67 | 11.11 | 66.67 | 3.67 | 7.44 | 3.67 | 0 | |
| 10 % | 7.45 | 8.03 | 0 | 0 | 3.67 | 3.67 | 66.67 | 7.44 | 3.67 | 0 | |
| 20% | 3.67 | 0 | 14.79 | 0 | 0 | 3.67 | 7.44 | 70.33 | 3.67 | 0 | |
| 35% | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 3.67 | 85.32 | 0 | |
| 100% | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 100 | |