Comparison of regression methods and Shaeffer’s formula in prediction of Live Body Weight of Ganjam Goats

Abstract

Linear regression and polynomial regression of order two and three were utilized to predict the live weight ofGanjam goats across five age groups using chest girth as predictor and their accuracies were compared with theprediction of weight made by Shaeffer’s formula. Live body weight of Ganjam goat recorded by electronic weighingbalance was used as standard for calculating the error of prediction. The body weights of 1014 Ganjam goats (329males and 685 females) were estimated by each technique during 2015 to 2017. Compared with electronic weighingscale, the body weight estimates in Ganjam goat exceeded in all age groups for Shaeffer’s formula whereas predicted body weight estimates by linear regression and second order polynomial regression were close to the live body weights. The estimates of linear regression and second order polynomial regression were significantly different from the electronic weighing scale for all age groups. The study concluded that polynomial regression of order two had better predictive value for live body weight of Ganjam goat, followed by third order polynomial regression, linear regression and Shaeffer’s formula, in order