Classification and Regression Tree (CRT) Analysis to Predict Body Weight of Potchefstroom Koekoek Laying Hens

摘要

Classification and regression tree analysis is a powerful statistical technique which helps to determine the most important variables in a particular dataset and helps to create a model. The study was conducted to identify linear body measurement traits (beak length, body length, keel length, chest circumference, toe length, body girth, shank length, back length, shank circumference and wing length) which could be employed in developing an effective prediction equation for body weight of Potchefstroom Koekoek laying hens. Eighty Potchefstroom Koekoek laying hens at twenty two weeks old were used. Pearson’s correlation together with classification and regression tree (CRT) methods were used for analysis. Descriptive statistics indicated that mean of body weight was 1.50 kg. Correlation findings revealed that body weight was positively significantly correlated (P 0.05) with beak length (r = 0.23) and toe length (r = 0.21), respectively. CRT results demonstrated that beak length, wing length and back length play an important role in the body weight of Potchefstroom Koekoek laying hen chickens. This study suggests that body weight of laying hens could be estimated by some linear body measurement traits. The models established in the current study might be employed by chicken farmers when making selection during breeding to improve body weight. However, further studies need to be done to validate the use of classification and regression tree analysis in prediction of body weight from linear body measurement traits of chickens.