Table 2.
Statistical performance of the linear discriminant analysis (LDA) models built for the differentiation of vinegars according to the PDO using the UV-Vis spectral data.
| Wavelength Interval | Preprocessing Method | Number of Components | Classification Performance | SENS | SPEC |
|---|---|---|---|---|---|
| 200–278 nm | Raw | 4 | 81.0% | 66.0% | 83.5% |
| SNV | 2 | 76.0% | 56.6% | 81.6% | |
| MSC | 2 | 77.0% | 60.3% | 82.9% | |
| 1D | 4 | 77.0% | 58.2% | 83.7% | |
| 2D | 7 | 77.5% | 45.6% | 74.5% | |
| 280–400 nm | Raw | 8 | 98.6% | 77.8% | 89.4% |
| SNV | 8 | 83.8% | 59.6% | 84.7% | |
| MSC | 9 | 83.0% | 61.1% | 85.7% | |
| 1D | 7 | 78.0% | 44.8% | 74.1% | |
| 2D | 3 | 68.3% | 43.5% | 70.8% | |
| 402–500 nm | Raw | 3 | 69.0% | 45.5% | 73.9% |
| SNV | 7 | 82.3% | 62.0% | 84.0% | |
| MSC | 7 | 80.0% | 59.1% | 84.0% | |
| 1D | 3 | 68.0% | 54.4% | 77.7% | |
| 2D | 6 | 71.8% | 37.5% | 70.6% | |
| 502–600 nm | Raw | 3 | 71.0% | 48.3% | 75.4% |
| SNV | 10 | 88.7% | 63.7% | 85.3% | |
| MSC | 11 | 85.9% | 61.7% | 84.6% | |
| 1D | 5 | 71.8% | 42.5% | 69.9% | |
| 2D | 4 | 67.6% | 42.2% | 70.9% | |
| 602–698 nm | Raw | 11 | 81.7% | 45.0% | 74.1% |
| SNV | 2 | 80.0% | 59.2% | 82.4% | |
| MSC | 5 | 80.0% | 60.7% | 83.8% | |
| 1D | 2 | 67.6% | 46.3% | 72.5% | |
| 2D | 2 | 66.9% | 41.3% | 70.0% |