CLUSTER ANALYSIS OF THE ESTIMATES FROM GROWTH CURVES

The main of this study was to assessment the use of nonlinear mixed model to growth using a data of weight-age simulated of pigs by QTLMAS2009 workshop. In addition, we also aimed to identify the groups of animals with the growth patterns similar over time. In the first step, the mixed nonlinear model, Gompertz, was fitted with parameters as random or fixed effect. Comparisons between the fixed and mixed effects models are made. In two-step, an analysis of cluster was applied with the estimate the parameters of each animal in order to recognize similar groups of animals. It was shown that the Gompertz model with random effect was suitable to explain the growth than nonlinear model with fixed effect. The cluster analysis of the parameters of the Gompertz discriminated animals into three groups. The statistical analysis methodologies used contributed to the analysis of the data. Keywords: Growth curves, longitudinal data, nonlinear models, swine.