Vergleich von Fixed- und Random-Regression Modellen bei verschiedenen Funktionsansätzen für Laktationskurven zur Vorhersage von Milchleistungen

Abstract

Mixed linear models can be used to improve the informational value of milk yield forecasts. For this purpose two different functional approaches for modelling lactation curves are as well compared as three linear mixed models with varying random effects of individual animals and lactation numbers. It can be shown that more complex random regression models fit significantly better than fixed regression models.