Extended Data Fig. 7: Wavelengths of importance within PLS-DA models.
From: Remote spectral detection of biodiversity effects on forest biomass
Wavelengths of importance in distinguishing species within partial least squares discriminant analysis (PLS-DA) models (red) shown alongside the PLSR model of stem biomass (grey, unchanged in all panels). The variable importance index represents the reduction of sums of squares49. Solid lines indicate the mean and shading indicates the 95% confidence intervals around the mean importance value for each wavelength across the 1000 model iterations. Vertical lines highlight the 20 most important wavelengths on average across the model iterations. Noisy and water absorption wavelengths are omitted.
