Using fuzzy data mining to evaluate survey data from olive grove cultivation

User knowledge about cultivation and soils is relevant in developing countries, given the lack of any other kind of information, and in developed countries where it may contribute to the sustainable and operational planning for farm production systems. But user data show a high degree of inaccuracy and uncertainty; hence the appropriate treatment of this kind of data in order to obtain potentially useful information requires suitable storage and processing techniques, such as fuzzy relational database and fuzzy data mining techniques. The user data were obtained from a survey with 126 variables carried out on 210 olive grove farms in the Province of Granada (southern Spain). A set of 34 variables were selected following a cleaning process, and 1420 fuzzy association rules relating handling, soil, or environment with olive fruit production (kg ha−1), % of oil in olive fruit or acidity of fruit on tree (olive oil quality), were obtained. Several of these rules can be considered as user evaluation rules (UER) because they show a clear association between the corresponding variables. It is demonstrated that the greatest influence of the variables for handling, soil and environment on the production and the quality of the olive oil, arises when those variables limit or reduce the production or quality. Some UER are corroborated with the knowledge regarding olive cultivation found in the bibliography. Some others contradict the knowledge and others reveal relationships not described previously. Because of this, the proposed working method can be considered as mostly exploratory. The final objective of this work is to help decision-making in the field of olive cultivation in Andalusia. However, the working method described in this paper is applicable to other geographical areas and to other kinds of crops.