Appraisal of growth curve in Sirohi goat using non-linear growth curve models.

Abstract

In the present study, the appropriateness of five non-linear mixed growth curve models was studied to describe body weight growth of Sirohi male and female goat from birth to 12months of age. Using selected functions, biologically pertinent variables were estimated for each goat from the Brody, Gompertz, Logistic, Richard, and Weibull. In addition, the males had higher asymptotic live weight 'A' than female counterparts. The highest values for 'B' parameter were observed for Weibull in both sexes, whereas the lowest value was calculated from Gompertz model. Maturity rate 'K' was higher in males than that of females. Male goat had heavier inflection weight 'm' and higher age than females for all the models. The best model was determined by considering the following goodness of fit criteria: coefficient of determination (R2), mean square error (RMSE), mean absolute error (MAE), and mean square absolute percentage error (MAPE). In both the sexes, all the non-linear regression models fitted the data well with high coefficient of determination (R2) ranging from 0.99 to 0.89 and 0.99 to 0.91 for male and female respectively. The Brody model provided the best fit of growth curve in both males and females and the lowest fit of growth were provided by Weibull model. We therefore, concluded that the Brody model was favorable for determining body weight in both sexes of Sirohi goat.