Abstract
Ranchers are continuously searching for suitable tools to rapidly and inexpensively assess the characteristics of donkey milk and because spectroscopic models are useful to assess the composition of many foods, an attempt to further improve the prediction performance of donkey milk protein, lactose and dry-matter content has been made using three widely used spectroscopic models by adding some interaction terms, namely product, ratio, sum and difference of absorbances for each couple of wavelengths. Principal component regression using product terms showed an improvement in prediction error achieving 1.8%, 1.7% and 0.9% for protein, lactose and dry-matter content respectively. Furthermore, the added ratio terms showed a very great improvement in the predictive overall performance achieving 0.3%, 0.4% and 0.2%. A coefficient has been found relating the widely used RPD, a standard index of prediction performance, to the new proposed “range of confident prediction error percent” being a more understandable parameter to assess the goodness of the prediction model.
© 2022 The Author(s)
PDF Article
More Like This
Cited By
You do not have subscription access to this journal. Cited by links are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.
Contact your librarian or system administrator
or
Login to access Optica Member Subscription
Add to BibSonomy