Data Mining Techniques for Modelling Seasonal Climate Effects on Grapevine Yield and Wine Quality

Abstract

The paper describes ongoing research in data mining techniques investigated for modelling seasonal climate effects on grapevine phenology that determines the ratio of grape berry composition that in turn determines the fineness of wine vintage in addition to winemaker experience and talent. A brief introduction to the literature in this problem domain is followed by a discussion on conventional statistical data analysis methods that looks at the problems in using these methods with only a decade old data, often considered as incomplete in sequence. Data relating to vineyard yield with its coincident seasonal climate change is used in this study to model seasonal climate effects at micro scales i.e., vineyard, using data mining techniques, decision trees and statistical methods. The initial results show potential for predicting future grapevine yield using vineyard data for more specific scenario building than is possible now, using macro climate data.