Use of multivariate analysis in the clustering of rural communities.

The identification and characterization of community clusters in terms of similarities and differences are important for the development and implementation of public policies focused on local realities, maximizing results and contributing to the increase of their effectiveness to the extent that communities of similar profile are identified according to their social structure. This paper aims at applying multivariate analysis to identify and create homogeneous groups of communities to be assisted by social programs and public policies. Specifically, the objective of this study was to typify, classify and characterize rural communities according to socioeconomic indicators and soil quality. Cluster analysis and graphical representation of Chernoff faces were used for this classification. Rural communities with similar characteristics were divided into three groups: Group 1: Ribeira (C1) and Jacundai (C2); Group 2: Sao Manoel (C3), Santa Maria do Mirindeua (C4), Santo Cristo (C5), Centro Ouro (C7), Sao Sebastiao (C8) and Santa Luzia do Tracuateua (C10); Group 3: Santana do Baixo (C6) and Santa Maria do Tracuateua (C9). The multivariate graph revealed that the communities present different profiles for the different variables in the socioeconomic dimension, confirming the clusters formed by some similarities within groups, as well as by heterogeneity between and within groups. Uso de analise multivariada no agrupamento de comunidades rurais Use of multivariate analysis in the clustering of rural communities Wilza da Silveira Pinto1* Waldenei Travassos de Queiroz1 Manoel Malheiros Tourinho1 Paulo de Tarso Eremita da Silva1 1Universidade Federal Rural da Amazonia – UFRA, Instituto de Ciencias Agrarias, Av. Presidente Tancredo Neves, 2501, Terra Firme, 66077-901, Belem, Para, Brasil Autor Correspondente: *E-mail: wilza.pinto@ufra.edu.br PALAVRAS-CHAVE Analise de agrupamento Faces de chernoff Grafico multidimensional