Study on Multiple Linear Regression and Principal Component Analysis for Prediction of Lifetime Performance of Kankrej Cattle

Abstract

The present investigation was conducted on 274 Kankrej cattle maintained at Livestock Research Station, Kodamdesar, Bikaner, calved  between 2012 to 2022 with the objectives to study principal component analysis (PCA) and multiple linear regression analysis (MLRA)  for prediction of lifetime performance of Kankrej cattle. Six early lactation traits (First lactation length- FLL, First lactation dry period FDP, First lactation 305 days’ milk yield- F305DMY, Second lactation length- SLL, Second lactation dry period- SDP, and Second lactation  305 days’ milk yield- S305DMY) were used to analyze the lifetime milk yield upto 5th and 7th lactations. The MLR analysis revealed that  the model containing F305DMY, SLL and S305DMY for upto 5th lactations’ lifetime milk yield, and model containing F305DMY, FDP  and SLL for upto 7th lactations’ lifetime milk yield having R2 = 68.3% and 68.5%, respectively, were found to be optimal models. PCA  revealed that the first 2 principal components (FLL, F305DMY) explained more than 78% of the total variation for LTMY5 and more than  81% variation for LTMY7. In this study F305DMY was found most important early trait in prediction of lifetime production of Kankrej  cattle on the basis of PCA and MLR analysis, out of which PCA was found to be better. Significant finding of this study may be helpful  in developing selection methodology for Kankrej cattle after validation in a large population.