Application of machine-learning methods to milk mid-infrared spectra for discrimination of cow milk from pasture or total mixed ration diets

The prevalence of “grass-fed” labeled food products on the market has increased in recent years, often commanding a premium price. To date, majority methods used for authentication grass-fed source are driven by auditing and inspection farm records. As such, ability to verify claims ensure consumer confidence will be important future. Mid-infrared (MIR) spectroscopy is widely dairy industry as rapid method routine monitoring individual herd milk composition quality. Further harnessing data from spectra offers promising readily implementable strategy authenticate at both processor levels. Herein, comprehensive comparison robustness, specificity, accuracy 11 machine-learning statistical analysis were tested discrimination versus non-grass-fed milks based MIR 4,320 samples collected cows pasture or indoor total mixed ration–based feeding systems over 3-yr period. Linear discriminant partial least squares (PLS-DA) demonstrated offer greatest level prediction cow diet spectra. Parsimonious strategies selection most discriminating wavelengths within also highlighted.