A large number of samples from different stages of the vinification process (from harvest to bottling) will be analyzed by classical (chemical) and organoleptic methods (tasting) as well as  by optical methods of analysis (eg UV-Vis, Raman, etc.) for different varieties and different winemaking operations. The results of chemical, organoleptic and optical analysis methods will be related using statistical methods. Thus, various organoleptic characteristics (e.g. color, sweetness, flavor, acidity, etc.) will be related to the chemical composition and spectral footprint of each wine by variety and wine-making operation, and algorithms will be established on how the particular parameters per vinification and variety affect the quality of the final product. Essentially, a "roadmap" for the production of quality wine will be created. By using these algorithms, the cooperating and other interested companies will be able to track the winemaking parameters that are of interest to them more easily and at a lower cost, to explore new winemaking protocols, to adapt the organoleptic characteristics of finished products to consumer requirements and generally to improve the quality of their products and thus increase their local and international market share.
In addition, natural antioxidants will be tested as part of the project instead of sulfides currently used in order to produce functional wines without chemical additives with increased added value for the producer as well as the consumer. Overall, the present proposal launches an industry and research partnership in the field of Agrofood, which will increase the added value of the wine sector in Greece and will improve its position as a winemaking country in Europe and internationally.
 

Application of Ultraviolet-Visible Absorption Spectroscopy with Machine Learning Techniques for the Classification of Cretan Wines

The present study was aimed at the identification, differentiation and characterization of red and white Cretan wines, which are described with Protected Geographical Indication (PGI), using ultraviolet–visible absorption spectroscopy. Specifically, the grape variety, the wine aging process and the role of barrel/container type were investigated. The combination of spectroscopic results with machine learning-based modelling demonstrated the use of absorption spectroscopy as a facile and low-cost technique in wine analysis. In this study, a clear discrimination among grape varieties was revealed. Moreover, a grouping of samples according to aging period and container type of maturation was accomplished, for the first time.