Phenotypic classification of sugarcane from near infrared spectra obtained directly from stalk using ordered predictors selection and partial least squares-discriminant analysis

Abstract

Author Summary: A new method was developed for the early selection of sugarcane genotypes using near infrared spectroscopy combined with partial least squares discriminant analysis (PLS-DA) and a variable selection method named ordered predictors selection (OPS). The use of the OPS method improved the predictive capacity of PLS-DA models to classify the sugarcane samples correctly according to fiber content (FC) and pol percent (PP).