In:
agriRxiv, CABI Publishing, Vol. 2022 ( 2022-01)
Abstract:
Traditionally, wine producers perform early wine quality prediction on-site based on the berry morphological and sensory characteristics together with the measurement of basic chemical parameters. Incorporating analysis on grape and wine volatiles could potentially achieve accurate prediction of wine quality, but forming these models requires careful selection of grapes, controlled fermentations and standardised quality assessment. Here, we present 3 models for the prediction of quality in Shiraz wine. Modelling was done by general regression analysis with 4-fold cross-validation. Model 1 (R 2 = 99.97% and 4-fold R 2 = 97.61%) for prediction of wine quality from wine volatiles, Model 2 (R 2 = 99.89% and 4-fold R 2 = 98.42%) for early prediction of wine quality from free- and glycosidically- bound grape volatiles, and Model 3 (R 2 = 91.62% and 4-fold R 2 = 80.21%) for prediction of wine quality from free grape volatiles only. The accuracy of these models presents an advancement in the early prediction of wine quality and provide a valuable tool to assist grape growers and winemakers in understanding quality in the vineyard to better direct scarce resources.
Type of Medium:
Online Resource
ISSN:
2791-1969
DOI:
10.31220/agriRxiv.2022.00162
Language:
English
Publisher:
CABI Publishing
Publication Date:
2022
