Aplicação da análise multivariada na caracterização dos municípios paranaenses segundo suas produções agrícolas
Resumo
Resumo: O objetivo geral deste trabalho é difundir a análise multivariada no estudo de dados agrícolas. Esse estudo trata da produtividade média de algumas culturas nos municípios paranaenses nas últimas safras. A partir da área e produção das culturas de algodão, milho, soja, milho safrinha, soja safrinha e trigo dos últimos anos, obtido junto à Secretaria de Estado da Agricultura e do Abastecimento – SEAB, Departamento de Economia Rural – DERAL, foi calculada através do Excel a produtividade média das diversas culturas por municípios das últimas safras. Foi realizada a estatística descritiva dessa produtividade média por culturas, através do Software Statistica. E novamente, através do Excel, construíram-se os gráficos de produtividade média das últimas safras. Dividiu-se o estado do Paraná em mesorregiões geográficas, de acordo com a divisão do Instituto Brasileiro de Geografia e Estatística – IBGE, para obtenção da produtividade média por mesorregião paranaense e construção dos gráficos. A partir da matriz de dados de produtividade média, por cultura e mesorregiões paranaenses, foi utilizada a Análise Multivariada, através da Análise Fatorial, com o Software Matlab R 12, que utiliza o método das componentes principais para estimar os pesos e as variâncias específicas e o critério de Kaiser. Essa análise permite identificar em ordem crescente as melhores mesorregiões paranaenses em termos de produtividade para cada cultura. Em complementação à Análise Fatorial, foi utilizada a Análise de Agrupamento (Cluster), com base na similaridade ou distância, ajustando-se o melhor algoritmo de correlação cofenética para a construção do dendrograma. Essa última análise permite agrupar as mesorregiões de acordo com uma espécie de distância. Os resultados mostraram que a análise fatorial e a análise de agrupamento se complementam entre si. As conclusões permitiram classificar e agrupar os municípios, que obtiveram as melhores médias de produtividades, através das mesorregiões paranaenses. Abstract: The general objective of this work is to show multivariate analysis in study of agricultural information. This study is about the average productivity of some agricultural products from the last harvests in Paraná State. The area and the production of the following crops we’ve given by the Agricultural Secretarist of Paraná – SEAB, and also by the Rural Economy Department – DERAL. Taking into consideration the area and the production of cotton, corn, soya, corn of autumn, soya of autumn and wheat in the last few years, it was calculated, using the Excel program, the average productivity of several products grown per town from the last harvests. The statistics of this average productivity was made by the Software Statistics; and using the Excel program again, the graphs of the average productivity from the last harvests were made. To do this, the state of Paraná was divided into geographical mesorregions, following the rules of the Brazilian Geographical and Statistics Institute – IBGE, in order to obtain the average productivity per mesorregions in Paraná and also in order to do the graphs. From the data of the average productivity, the Multivariate Analysis was made using the Factor Analysis with the Software Matlab R 12. This software is able to use the method of the main components to estimate the weight, the specific changes and the Kaiser criterium. This analysis points in a crescent order, the best mesorregions in Paraná State in terms of productivity for each agricultural product. And to complement the Factor Analysis, it was used the Cluster Analysis, based on the similarity or distance, fitting the best algorithm of cophenet correlation for the construction of the dendrogram. This last analysis allows grouping the mesorregions according to a type of distance. The results show that factor analysis and cluster analysis are complementary. The conclusions let classify and allows to groups the municipality that obtain the better average of productivity through the mesorregions in Paraná State.
Collections
- Teses & Dissertações [9328]