comparison of the estimators obtained from ordinary least squares and principle component analysis methods to predict carcass yield in lori goats breed

the present study was carried out to establish the relationships between carcass yield and some body measurements in lori goat breed. the first body weight (bw) and morphological traits of 186 heads goat were recorded. the animals were then slaughtered to calculate the carcass yield. results showed the colinearity among the traits. in order to eliminate colinearity problems, principal component analysis was used. the final predicted coefficients of carcass yield for carcass weight, body weight, heart girth, paunch girth, height at wither, body lengthbl, animal length and body condition score was 0.0049, 0.0006, -0.0016, -0.0029, -0.0008, 0.0008, 0.0001 and 0.0175, respectively. results showed that the problem of multicollinearity in the relationship between carcass yield and independent variables in goats can be solved using principal component analysis. this method leads to more stable and reliable coefficients with less standard error than those from ordinary least squares. furthermore, body measurements can be used as selection criterion to improve carcass yield that it cannot be recorded on a live animal.