Josefina Barrera Morelli: Predicting The Milkywhey

<p dir="ltr">This study investigated the effects of spectral preprocessing and regression algorithms for the prediction of individual fatty acids (FA) using MIR, particularly assessing whether ML techniques can improve models from the standard approach, PLSR. </p><p dir="ltr">A 1,242 bulk milk samples with concentrations for 25 individual FA determined by GCMS and MIR spectral data. Spectra was preprocessed by removing the regions of no interest and used, along with the analytical concentrations to develop calibration models with PLSR and four machine learning algorithms.</p>