The study was undertaken to determine the sex and age specific best fitted regression model for prediction of live weight in Surti kids under field conditions. A total of 1151 observations on 544 male and 607 female kids of different age groups were taken randomly. Seven prediction equations used in regression analysis involving all possible combinations of body measurements and weight was regressed on body measurements. Based on R2 values, sex dependent (S) regression equation were found to be better predictor of body weight in Surti kids. However, when equations were further cross verified by comparing predicted weight with actual weight using correlation coefficient, mean paired difference and range of paired differences at 95% confidence interval, age dependent (A) regression equations were found to be best fitted. Finally, 12 best fitted regression equations were obtained in which length-girth combination and girth alone were found to be better predictor of body weight. It was ascertained that R2 alone does not make a good indicator of prediction accuracy of the model and should be cross verified using other available statistical tools.
Body weight prediction, Goat, Regression model, Surti
