Skip to main content

Table 2 Validation and independent validation statistics as well as number of samples used for NIRS model development

From: Nutrient retention and release in coarse woody debris of three important central European tree species and the use of NIRS to determine deadwood chemical properties

Variable Validation set Independent validation set Number of samples (calibration/validation/independent validation)
R 2 RMSEP R 2 RMSEP RPD Ranking according to Malley et al. (2004)
C 0.89 0.879 0.79 0.901 2.17 D 140/151/576
N 0.94 0.042 0.94 0.041 4.12 A/B+ 135/146/585
P 0.87 0.0278 0.87 0.034 2.93 C 48/35/20
Lignin 0.99 1.14 0.98 1.55 6.79 A 48/36/22
Extractives 0.87 1.05 0.92 0.76 3.72 B 84*/22
  1. *Crossvalidation; + Models were ranked “A” according to RPD value, but “B” in relation to R2