Self-Organizing Map approach to cluster Brazilian agricultural spatiotemporal diversity

This work aims to cluster Brazilian municipalities according to their spatiotemporal agricultural diversity pattern. The diversity index has been defined for eight categories and calculated by Shannon’s entropy index from annual (1999-2018) IBGE’s estimates for agricultural production. The proposed clustering method is based on the Self-Organizing Map, an unsupervised artificial neural network, and comprises visual and automatic steps. The method partitioned the municipalities into eight groups spatially organized in three regions showing different spatiotemporal patterns.