Proposal for a non-linear model applied to sheep growth data

The main goal of this work is to propose a non-linear model. It also aimed to compare among the Proposed model, the Logistic, Gompertz, von Bertalanffy, Brody and Richards models, the one that presents the highest quality of fit for the description of Île de France sheep breed. To do so, data from male sheep raised in confinement were used. The parameters of each model were estimated by the least squares method using the iterative Gauss-Newton process. The quality of fit of the models to the data was verified through the criterion of the adjusted coefficient of determination (), mean square of the residue (MSR), Akaike's information criterion (AIC) and mean absolute deviation (MAD). In addition to the results of the different quality of fit criteria, cluster analysis was also used to classify and select nonlinear growth models. Among the adjusted models, only the Richards model did not show convergence for the sheep weight. The Proposed and Gompertz model presented the best quality of fit evaluators, superior to the other models to describe the Île de France sheep growth curve.