Comparison of Non-Liner Growth Models to Describe the Growth Curve from Birth to Yearling in Markhoz Goat

The objective of this study was to select the best model among five non-linear growth functions, i.e., Brody, Gompertz, Logistic, Von Bertalanffy and Negative exponential for describing the growth curve in Markhoz goat. The data included 5557 body weight records of goats from birth to yearling which were collected during 2006 to 2013 at Sanandaj Research Station. Growth curve parameters (A, B, K) were estimated by the NLIN procedure of SAS program and in order to compare different models for selecting the best model, statistics of coefficient of determination (R2adj), Akaike´s Information Criterion (AIC), Bayesian Information Criterion (BIC), Root Mean Squares Error (RMSE) and Durbin-Watson were calculated. For all animals, Logistic model was the best model as provided the greatest R2adj and the lowest values of RMSE, AIC and BIC than other models and this model could indicate the best prediction of growth in Markhoz goat. The best selected models for predicting growth of males and females were Brody and Gompertz models, respectively; and Logistic model was selected as the best model for both singles and twins. The negative exponential model provided the worst fit of growth curve for all animals, males, females, singles and twins. Logistic model can provide the best prediction of growth pattern of Markhoz goat over a definite time period.