Table 2. The best model resulting from CV-PLSR analysis
| Density at MC | Part surface | Pre-processing spectra treatment | N | Number of LV | R2C | R2CV | RMSEC (g/cm3) | RMSECV (g/cm3) | RPD |
| Green | Transverse | SNV | 65 | 7 | 0.4650 | 0.2015 | 0.0531 | 0.0648 | 1.1950 |
| Bark | MSC | 65 | 10 | 0.6488 | 0.3508 | 0.0418 | 0.0569 | 1.3617 |
| Air-dry | Transverse | 1st der. with 25 sp | 65 | 7 | 0.5949 | 0.3472 | 0.0555 | 0.0705 | 1.3216 |
| Bark | 1st der. with 25 sp | 65 | 12 | 0.8852 | 0.5154 | 0.0282 | 0.0580 | 1.6061 |
CV-PLSR: cross-validation partial least squares regression, MC: moisture content, LV: latent variable, RMSEC: root-mean-square error for calibration, RMSECV: root-mean-square error for cross-validation, RPD: ratio of performance to deviation, 1st der.: first derivative, sp: smoothing points.