Animal growth models for irregular times

Abstract:

The objective of this work was to propose models that consider an irregular data structure and to evaluate them in relation to models used for regular times. The Gompertz, Logistic, and Von Bertalanffy growth models, with regular and irregular structures for the errors, were considered. The methodology was exemplified using real and simulated data. Sixteen average weights of 160 animals of the Hereford breed were used, with weighing performed from birth up to approximately 2 years of age. For each model, the parameters of the best adjustment were used for the simulation. The model adjustments improve when the original structure of the data is considered, reducing the sum of the squares of the residues and reducing the value of the Akaike criterion.

Index terms:
autocorrelation; weight gain; Hereford; nonlinear models.