GAMLSS na experimentação agropecuária: um estudo em doenças parasitárias de bovinos de leite

• Autor(a) principal: Roquim, Fernanda Venturato
• Data de Publicação: 2018
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Institucional da UFLA
• Texto Completo: http://repositorio.ufla.br/jspui/handle/1/29740
• topic: Bezerras leiteiras; Excesso de zeros; Dados de contagem; Experimentação agropecuária; Agricultural and livestock experimentation; Count data; Dairy cattle; Inflated zeros; Estatística

This study shows the main characteristics of the generalized additive models for location, scale and shape (GAMLSS). Furthermore, describes its estimation methods, which includes the iterative algorithms, the inference and prediction methods, briefly discussing possible distributions and additive terms and, finally, model selection and diagnostics techniques. The emphasis was to present the flexibility of such class of models which allows us to fit all parameters of the response variable distribution as a function of explanatory variables. In GAMLSS, a wide range of distributions are possible to explain the response variable with the possibility of incorporating smoothing functions, which offer models that are able to increasingly portray reality. The aim of this study was to demonstrate how this models can be used in agricultural and livestock experimental data, bringing a practical application related to the study of treatments for parasitosis in dairy cattle. In addition, it was presented some papers available in the scientific community that have already used the GAMLSS within this area.