Análise exploratória de dados de área utilizando o R

• Ano de defesa: 2018
• Autor(a) principal: Almeida, Pablo Lourenço Ribeiro de
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Lavras; Programa de Pós-graduação em Estatística e Experimentação Agropecuária; UFLA; brasil; Departamento de Estatística
• Palavras-chave em Português: Estatística espacial; Autocorrelação espacial; Detecção de clusters; Spatial statistics; Spatial autocorrelation; Cluster detection; Estatística
• Link de acesso: http://repositorio.ufla.br/jspui/handle/1/36979

The present work aims to produce a material describing in detail the spatial data analysis on lattices by considering global autocorrelation indices, local autocorrelation and clusters detection tests. The methods studied in this work were applied in a dataset of classical dengue fever from of the city of Campina Grande, PB, Brazil, obtained from the Municipal Health Office. The statistical analyzes were performed using the statistical softwares R and SaTScan. It was observed that there were global and local spatial autocorrelation for the incidence of the dengue fever. Furthermore, it was compared the power of detecting risk areas by using maps with the information about the local indication of spatial association (LISA) and spatial scan statistics. It was concluded the LISA presented smaller power than the scan statistics to detect the districts with highest risk of dengue fever transmission.