Application Problem

Farmers want to know what the best areas of their farmland are to grow soybean. Thus, the main goal of this application is to make recommendations to farmers that want to know what the best farmland areas are to grow soybean. Since this type of recommendation is very complex and has to consider several variables (e.g., soil features, climatology, economic factors), we here only consider the following soil features and its respective characterizations:

• Texture - a texture can be sandy, medium, clayey, and very clayey.
• Drainage - a drainage can be slow, moderate, high, and excessive.
• Organic Matter - an organic matter can be low, medium, high, and extremely high.

By following the FIFUS definition, each soil feature composes a Fuzzy Spatial Antecedent (FSA), where each characterization is represented by a fuzzy spatial object. For instance, a sandy texture is represented by a fuzzy region stored in the spatial database. The Consequent Set of this application is the recommendation level to cultivate soybean in a given location (i.e., point, line, or region). In this problem, the inference algorithm process 48 fuzzy rules in order to return an inferred value.