Aplicação da teoria de modelos mistos a dados longitudinais

• Ano de defesa: 2016
• Autor(a) principal: Simão, Sérgio Domingos
• 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 Ciências Exatas
• Palavras-chave em Português: Medida repetida; Parcela subdividida; Estrutura de covariância; Análise de resíduos; Repeated measure; Estatística
• Link de acesso: http://repositorio.ufla.br/jspui/handle/1/12007

This study aims to develop a material describing in detail the analysis of longitudinal data using linear mixed mo dels and addressing waste analysis. We p erformed an exploratory analysis of the data, the comparison of different structures of variance-covariance and comparison of share of results sub divided in time with the analysis of mixed mo dels for longitudinal data. Statistical analyzes were carried out using the statistical program R and with the SAS statistical software. The choice of the b est covariance structure will p erformed considering the Akaike information criterion (AIC) and Bayesian criterion Schwarz (BIC). The variables used for illustration of the analysis were: average feed intake (g.bird − 1.day − 1) and yolk weight (grams). For the variable average feed intake was chosen covariance structure unstructured, while for the variable yolk weight used a first-order autoregressive structure (AR (1)). This manual is useful for researchers, teachers and students of applied fields, to perform the analysis of longitudinal data using linear mixed models considering different covariance structures and contemplating the residuals analysis. The choice of covariance structure affected the significance of the F test of fixed effects for the variable average feed intake. For the variable yolk weight the choice of the structure did not affect the significance of the F test. The residuals analysis was important in the verification of the assumptions of the model and the general characteristics of the data.