Regression model analysis for prediction of body weight from body measurements in female Batur sheep of Banjarnegara District, Indonesia

Ibrahim A, Artama WT, Budisatria IGS, Yuniawan R, Atmoko BA, Widayanti R. 2021. Regression model analysis for
prediction of body weight from body measurements in female Batur sheep of Banjarnegara District, Indonesia. Biodiversitas 22: 2723-2730. Bodyweight is an important aspect of livestock management. The present study was undertaken to estimate correlation coefficients between biometric traits and identify best predictor of body weight in female Batur sheep from body measurements. Data on
body weight and body measurements (body length: BL, chest girth: CG and withers height: WH) were collected from 73 female Batur sheep in Batur Village, Banjarnegara District, Central Java Province, Indonesia. Batur sheep were grouped into 3 categories based on their age, namely groups <1.5 years, 1.5-2.5 years and >2.5 years. The data were analyzed using simple, multiple, and automatic linear
regression methods using the SPSS computer software version 25 platform. The correlation coefficient, coefficient determination,
adjusted coefficient determination, residual standard error, Akaike information criterion, Bayesian information criterion, and Akaike information criterion corrected were used to determine the best regression formula for the prediction of BW. The average BW (kg), BL (cm), CG (cm), and WH (cm) of 49.27, 63.11, 91.41, and 56.82, respectively was observed in the present study. The correlation coefficients of 0.433, 0.866, and 0.369 for BW with BL, CG, and WH were observed in the present study. The best prediction of BW
using two predictors (BL and GC) was BW =-56.522 + 0.509BL + 0.843CG, followed by using three predictors (BL, CG, and WH) was BW =-57.897+ 0.505BL + 0.839CG + 0.034WH, and using the only one predictor (CG) was BW =-28.443 + 0.905CG. The study
revealed that CG and its combination with other linear body measurements can effectively define the body weight in Batur sheep. However, the highest R2 of 0.782 was observed when CG and BL were used as predictors.