Determination of the Some Body Measurements Effecting Fattening Final Live Weight of Cattle by the Regression Tree Analysis

Authors : Rifat AKŞAHAN, İsmail KESKİN

Pages : 53-59

View : 18 | Download : 32

Publication Date : 2015-06-25

Article Type : Research Paper

Abstract :The aim of this study is to determine body measurements affecting final live weight insert ignore into journalissuearticles values(FLW); at fattening period of 103 young insert ignore into journalissuearticles values(Holstein, Simmental, Brown Swiss and crossbreed); bulls reared in Bolvadin district of Afyon province of Turkey. For this aim, a regression tree method on the basis of CHAID insert ignore into journalissuearticles values(Chi-Squared Automatic Interaction Detector); algorithm were used. FLW was considered as response variable. The independent variables included in the study were withers height insert ignore into journalissuearticles values(WH);, back height insert ignore into journalissuearticles values(BH);, front rump height insert ignore into journalissuearticles values(FRH);, back rump height insert ignore into journalissuearticles values(BRH);, body length insert ignore into journalissuearticles values(BL);, back rump width insert ignore into journalissuearticles values(BRW);, chest dept insert ignore into journalissuearticles values(CD);, chest circumference insert ignore into journalissuearticles values(CC);, and fattening period insert ignore into journalissuearticles values(FP);. Minimum numbers of young bulls in parent and child nodes for CHAID algorithm were assigned as: 8:4, in order to obtain the best model quality criteria for the data evaluated under the study. Model quality criteria like coefficient of determination insert ignore into journalissuearticles values(%);, adjusted coefficient of determination insert ignore into journalissuearticles values(%);, and the Pearson correlation between actual and predicted fattening FLW values were estimated to determine the effectiveness of CHAID algorithm. The independent variables influencing FLW were CC insert ignore into journalissuearticles values(Adj. P 190 cm, averagely. The secondly heaviest sub group of the young bulls with 183 4 months provided a FLW of 505.562 kg. Consequently, the use of CHAID algorithm for predicting FLW from various body measurements could be advised to be useful for breeding pur-poses.
Keywords : CHAID algoritması, Regresyon Ağacı, Siyah Alaca, Simental, Esmer Sığır