Times are displayed in (UTC-04:00) Eastern Time (US & Canada)Change
Exploring Wine Terroir with Machine Learning
Topics: Wine, Beer, and Spirits
, Spatial Analysis & Modeling
,
Keywords: wine, terroir, machine learning Session Type: Virtual Guided Poster Abstract Day: Monday Session Start / End Time: 2/28/2022 11:20 AM (Eastern Time (US & Canada)) - 2/28/2022 12:40 PM (Eastern Time (US & Canada)) Room: Virtual 49
Authors:
Jamison Conley, West Virginia University
Lee Ann Nolan, West Virginia University
,
,
,
,
,
,
,
,
Abstract
The concept of terroir came from the wine industry, with the idea that you can taste the place—that characteristics of a location, like its climate and soil, have notable impacts on the flavor of the wine. While widely regarded as a true phenomenon, quantitative studies evaluating terroir are rare. This research uses machine learning to evaluate the extent of terroir in wine tasting notes. Descriptions of wines are taken from Wine Spectator magazine, and a series of machine learning and statistical classifiers, including neural networks, random forest, and categorical regression, are employed to use the descriptions to predict the country where the wine was grown. Results support the expectations that the presence of terroir ensures that there are distinct country-scale flavor profiles that allow one to taste the place where the wine was grown.
