PREDICTING BODY WEIGHT FROM BODY AND TESTICULAR CHARACTERISTICS OF BALOCHI MALE SHEEP IN PAKISTAN USING DIFFERENT STATISTICAL ANALYSES

M. Jahan, M. M. Tariq*, M. A. Kakar**, E. Eyduran*** and A. Waheed****

Livestock and Dairy Development Department Balochistan, Quetta, Pakistan
*Center for Advanced Studies in Vaccinology and Biotechnology (CASVAB), University of Balochistan, Quetta, Balochistan, Pakistan.
**Balochistan University of Information Technology, Engineering and Management Science, Quetta
***Biometry Genetics Unit, Department of Animal Science, Faculty of Agriculture, Iğdır University, 76000, Iğdır-Turkiye.
****Faculty of Veterinary Sciences, Bahauddin Zakariya University, Multan, Pakistan
Mohammad tariq <tariqkianiraja@hotmail.com>

ABSTRACT

The aim of this study was to determine the suitable statistical analysis for the prediction of body weight from biometrical and testicular traits in Balochi male sheep. For this aim, statistical performances of Stepwise Regression Analysis, use of factor analysis scores with multiple regression model analysis and Ridge Regression analysis were evaluated on data of 131 Balochi male sheep. The measured characteristics were: body weight (BW), testicular length (TL), scrotal length (SL), scrotal circumference (SC), body length (BL), withers height (WH) and heart girth (HG). In order to determine the best model, determination coefficient (R2%), Root of Mean Square Error (RMSE) and Variance Inflation Factor (VIF) values were used. Stepwise Regression and Ridge Regression analyses produced multicollinearity problem due to high VIF and RMSE values. In comparison with these two analyses, use of factor analysis scores with multiple regression model analysis offering optimal solution with very low VIF and RMSE values was adopted for the prediction of body weight of the Balochi male sheep. In the factor analysis, the 3 new-uncorrelated variables derived from eight explanatory variables were used as explanatory variables with the multiple regression analysis. Results reflected without multicollinearity problem that 91.1 % of variation in body weight was perfectly explained by the 3 new uncorrelated variables.
Key words: Balochi Sheep, Factor Analysis, Ridge Regression, Multicollinearity, Multiple Regression, Stepwise Regression.