Matlab tutorial for spectroscopists. Part 8: partial least squares regression

I n this issue we will have a look at how you can do partial least squares regression (PLSR) in Matlab. The dataset is the same we used in part 7 of this series, and consists of 150 samples of ground beef, with fat levels ranging from 2% to 73%. The spectra were recorded in reflectance mode using standard sample cups on a NIRSystems 6500 instrument. All samples were recorded at five temperature steps between –5°C and +3°C. Only the +3°C spectra will be used in the present article. Reference data for moisture and protein are also available for these samples, but will not be used here.