Precision viticulture data analysis using fuzzy inference systems

Abstract

Aims: Various types of data are likely to be used in a precision viticulture framework, to adjust management actions according to within field variations. This paper proposes an alternative way of analysis to classical methods. Methods and Results: Data are analysed using fuzzy logic techniques. The result is a set of linguistic fuzzy rules induced from data. In this paper, the rules are build in order to explain the relationship between vintage quality, reduced to sugar content, and other available variables. The resulting system is proved to be accurate, moreover thanks to fuzzy logic interpretability, the induced rules are analyzed and compared to expert knowledge. Conclusion: This example highlights the potential of fuzzy logic to deal with precision viticulture datasets. Significance and impact of study: This is a preliminary work, it has been carried out using a free software available in the internet.

Keywords

• Precision viticulture
• machine-learning
• fuzzy logic
• nonlinearity
• interpretability
• linguistic models
• spatialized data
• Fispro