Regression tree analysis for predicting slaughter weight in broilers

Abstract

In this study, Regression Tree Analysis (RTA) was used to predict and to determine the most important variables in predicting the slaughter weight of Ross 308 broiler chickens. Data for this study came from 224 chickens raised during three different seasons, namely spring (n=66), summer (n=66), winter (n=92). Second week body weight, shank length, shank width, breast bone length, breast width, breast circumference and body length were used to predict the slaughter weight. Results of RTA showed that among the seven independent variables only four were selected, namely; body weight, breast bone length, shank width, and breast circumference. These selected independent variables were more efficient than the others in predicting the slaughter weight.RTA indicated that the birds which had values of second week body weight >295.95 g, breast bone length >55.82 mm and breast circumference >14.18 cm or that of body weight ≤295.95 g, breast bone length >60.26 mm and shank width >8.32 mm could be expected to have higher slaughter weights.