PRINCIPAL COMPONENT REGRESSION ANALYSIS (PCRS)- AN APPROACH FOR LIFETIME MILK YIELD PREDICTION IN MURRAH BUFFALOES

Abstract

Principal Component Regression Analysis (PCRA) has been carried out to formulate lifetime milk yield prediction model with principal components as predictors. Principal compoents (PCs) were based on initially expressed reproductive traits and part lactation records from 10 years data (1-4-1999 to 31-3-2009) of Murrah buffaloes, of Cattle of Buffalo farm of IVRI. Models for lifetime milk yield LTMY4 (lifetime milk yield as total milk yield up to four lactations) has been evolved with retained four PCs (explained 86.67% variation of original data). Seven types of model have also been fitted to have best model for LTLMY4 with first PC_1 as predictor. The model for LTLMY4 could explained 43.30% variation in the estimated values with adjusted R2=39.52%. Curve estimation analysis shows appropriateness of the Logarithmic function (adjusted R2-value- 43.05%) followed by Cob Douglas function (42.89%).