Aplicação do software R na classificação de áreas associadas à produtividade da soja e variáveis agrometeorológicas por meio de agrupamento Fuzzy

Abstract

This study aimed to apply an approach on the use of the R software based on fuzzy clustering for classification of areas associated with soybean yield, together with meteorological variables: rainfall average, average air temperature and solar radiation average. The study was conducted involving 50 municipalities in the western region of the state of Paraná, Brazil, with data from the harvest of the agricultural year 2010/2011. Through the algorithm Fuzzy C-Means, it was possible to form groups of municipalities similar to soybean yield, the identification of the appropriate number of clusters was obtained using the Entropy Partition Modified (MPE), the groups were validated by the Method Decision by Major degree of Pertinence (MDMGP). To measure the level of similarity of each cluster, we used the Index of Cluster Similarity (ISCl). Within the perspective of this study, the method used was adequate, allowing to identify clusters of municipalities with degrees of similarities in the order of 66 to 68%.