BERNSTEIN POLYNOMIAL APPROACH AGAINST TO SOME FREQUENTLY USED GROWTH CURVE MODELS ON ANIMAL DATA

Abstract

Non-linear Logistic, Gompertz and Richards growth curve models were fitted to the data from Simmental x Southern Anatolian Red (SAR) crossbred cattles. Individual growth curves were fitted based on live weight measurements, and then general growth curves were obtained for all the models. In addition, Bernstein basis polynomials have played important roles in nonparametric curve estimation. Therefore, Bernstein polynomial approach was used to model the growth curve in the current data. We determined the accuracy of the models by using coefficient of determination (R-2), Mean square error (MSE) and iteration number together. In summary, the most suitable model based on the accuracy criteria was Bernstein model. Among the well-known growth curves, Logistic, Richards and Gompertz were ordered respectively.