NONLINEAR REGRESSION MODELS AND APPLICATIONS IN ANIMAL SCIENCE

Abstract

Time dependent data, such as growth, yield and production in animal husbandry, are often expressed in nonlinear models. For this reason, nonlinear regression models take an important role in animal science. In this study, general information about nonlinear regression models was given and examples related to application areas in animal husbandry were referred. Nonlinear models that best describe the lactation curves, which show the change of milk yield with time, the growth curves of the animals, the egg weight and the production curves, were investigated. For correct model selection, after application of different models, criteria such as determination coefficient, adjusted determination coefficient, mean square error, Wellmot agreement criterion, absolute percentage error, Akaike information criterion, Bayesian information criterion were introduced. In addition, the steps for solving the models in SAS, STATISTICA, SPSS, MINITAB statistical package programs have been shown. As a result, the most used models were selected and a base of nonlinear models applied in animal science was created.