Multi-factor analysis and modeling of net energy of lactation (NEL) prediction in primiparous dairy cows

Abstract

Abstract In order to predict the net energy of lactation primiparous dairy cows, a prediction model was established. Body weight, milk yield, milk composition and feed nutrients were measured from 1266 Holstein primiparous dairy cows. Eight variables as lactation month (LM), milk yield (MY), milk fat content (MFC), milk protein content (MPC), milk lactose content (MLC), ratio of MPC to MFC (PF), ratio of MLC to MFC (LF), and ratio of MLC to MPC (LP) were set for prediction. The prediction model was built, tested and verified using correlation analysis, least squares method and LASSO multiple linear regression analysis. The results showed that there were five variables, Square (LM), Exp (MY), MPC, PF and LP, that had linear correlation with the net lactation energy of every next month. The LASSO multiple linear regression analysis simplified the prediction variables, and obtained a prediction model based on lactation time and LP, which enabled a simple, convenient and fast prediction of the net lactation energy of primiparous dairy cows for next month.