Use of genetic algorithm on mid-infrared spectrometric data: application to estimate the fatty acids profile of goat milk

Abstract

To know and to control the fine milk composition is an important concern in the dairy industry. The mid-infrared (MIR) spectrometry method appears to be a good, fast and cheap method for assessing milk fatty acid profile with accuracy. Although partial least squares (PLS) regression is a very useful and powerful method to determine fine milk composition from spectra, the estimations are often less accurate on new samples coming from different spectrometers. Therefore a genetic algorithm (GA) combined with a PLS was used to produce models with a reduced number of wavelengths and a better accuracy. Number of wavelengths to consider is reduced substantially by 5 or 10 according the number of steps in the genetic algorithm. The accuracy is increased on average by 9% for fatty acids of interest.