Variable Selection on Reflectance NIR Spectra for the Prediction of TSS in Intact Berries of Thompson Seedless Grapes

Fourier-transform near infrared (FT-NIR) reflection spectra of intact berries the grape variety Thompson seedless were used to predict total soluble solids (TSS) content. From an initial dataset, 12 subsets considered by applying variable selection extract reflectance values at wavenumbers most correlated chemometrically measured TSS The datasets processed both multiple linear regression (MLR) and partial least squares (PLS) methods towards predicting content from each spectrum. Prediction accuracy was in terms coefficient determination R2 root mean squared error (RMSE). It found that improved prediction with processing methods; up 0.972 0.926 RMSE 0.306 0.472 reported MLR PLS, respectively. combination displayed (a) higher when dataset variation limited, (b) lower large such as those a maturity stages, (c) failed more complex sets different harvest years. PLS has demonstrated reliable results various degrees complexity.