Técnica de agrupamento na seleção de modelos de regressão não lineares para descrição do acúmulo de matéria seca em plantas de alho

• Ano de defesa: 2014
• Autor(a) principal: Puiatti, Guilherme Alves
• Orientador(a): Cecon, Paulo Roberto
• Banca de defesa: Ferreira, Adésio
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Viçosa
• Programa de Pós-Graduação: Mestrado em Estatística Aplicada e Biometria
• Departamento: Estatística Aplicada e Biometria
• País: BR
• Palavras-chave em Português: Análise de agrupamento; Alho; Análise de regressão; Allium sativum
• Palavras-chave em Inglês: Cluster analysis; Garlic; Regression analysis; Allium sativum
• Área do conhecimento CNPq: CNPQ::CIENCIAS AGRARIAS
• Link de acesso: http://locus.ufv.br/handle/123456789/4078

Studies of genetic divergence between individuals or populations of plants and its growth path are of great importance in improvement programs, being essential for an appropriate handling through relevant information. Cluster analysis and regression models are techniques largely employed for such. The objective of this study was to identify and to group nonlinear regression models that best describe the dry matter accumulation of garlic plants over time (60, 90, 120 and 150 days after planting). 15 garlic accessions belonging to the Vegetable Germplasm Bank of Universidade Federal de Viçosa (BGH/UFV) were selected. The following nonlinear regression models were adjusted for each one of the accesses: Brody, Gompertz, Logistic, Mitscherlich and von Bertalanffy. The fit quality of the models was measured by the determination coefficient ( R 2 ); mean squared error ( MSE ); mean absolute deviation of the error ( DMA ); Akaike information criterion ( AIC ); and Bayesian information criterion ( BIC ).Then, for each access, the models were submitted the grouping analysis, with the appraisers of fit quality used as variables, using the UPGMA algorithm, the Mahalanobis distance as dissimilarity measure, and Mojena's method to obtain the number of groups. The adjusted models for each access were again grouped following the same criterion, but using the parameters with biological interpretation as variables, and the results of the different clustering analysis were then confronted. Comparing the results, it was observed that the Gompertz, Logistic and von Bertalanffy models presented better results, and they had close results in fit quality and estimates of the parameters. These three models, especially the Logistic, were efficient for the description of dry matter accumulation of garlic plants.