Fuzzy logic tool for wine quality classification

Abstract

Wine quality is a complex attribute, thus single berry parameters are not adequate to define grape suitability for the production of specific wines. Fuzzy logic systems are particularly suited for aggregating multiple data to feed multi-variables decision support systems. The aim of the present study was the development of a simple in use and reliable tool based on fuzzy logic multi criteria decision making to objectively classify wine quality based on selected grape attributes. For this purpose, representative berry samples were harvested and vinified for two consecutive years from thirteen commercial vineyards in Nemea, Greece planted with Vitis vinifera cv. Agiorgitiko. Total soluble solids, pH, berry volume, botrytis infection, grape seed colorization, anthocyanin extractability, optical density (OD 520) and skin phenolics (Dpell) were measured at harvest and were used in the tool as inputs. Moreover, the produced wines were sensory evaluated by an experienced and trained panel. The ranking of the vineyards, according to the tasting panel, was compared to the ranking made by the tool and the results showed high general agreement between them suggesting that the latter was able to model expert knowledge successfully. According to the results, the fuzzy logic multi criteria decision making tool could allow the incorporation of grape quality parameters at harvest into a single index providing grape growers and wine producers with a valuable tool for classifying wine quality.