Statistical exponential models adjustment to the growing of Nelore cattle. Adjustment and computational aspects

Abstract

Weight records of 697 Nelore animals were used to fit four non-linear modals (Brody, Bertalanffy, Gompertz and Logistic models) to Nelore cattle growth. Taylor's series with the "internal least squares" technique was used to estimate the parameters for each of the four non-linear models. These models were approximately equal to 99%. Brody's model had the slowest convergency and in this model many of the parameter estimates had no biological importance for interpretation. The number of weight records of each animal as well as the animal's latest age or weight had the highest influence in the fitting. In addition, changes in the animal growth index, mainly at later stages, resulted in computational difficulties and less precise parameter estimates.