Description of Santa Ines sheep growth using non-linear models selected by multivariate analysis

Abstract

The objective of this study was to describe the growth of Santa Ines sheep using nonlinear models selected by multivariate analysis. Seventeen nonlinear models were used to describe the growth of sheep from birth to 210 days of age. The models were classified by means of multivariate analysis using the adjusted coefficient of determination, the Akaike information criterion, the Bayesian information criterion, the mean squared error of prediction, the coefficient of determination, mean absolute deviation and percentage of convergence of prediction as evaluators of adjustment quality. The Richards, Reynolds and Weibull models did not converge and the Schnute, Meloun II, IV and VI, Gamito, Michaelis Menten and Modified Michaelis Menten models did not show any traditional growth pattern. Models Mitscherlich, Gompertz, Logistic, Meloun I, III and V, Brody and Von Bertalanffy converged and were used in grouping analysis. Of the four formed clusters formed, the first, composed by models Mitscherlich, Meloun I and Brody, showed the best values for all the evaluators considered. The curve parameters were influenced by gender and type of birth (P < 0.05) and the inflection point occurred at 94 days of age. The Mitscherlich, Meloun I and Brody models are adequate and equivalent to describe the growth of Santa Ines sheep.