Bayesian Hierarchical Modeling Applied to Test Day Milk Yield Data

Abstract

Fitting non-linear models to the milk yield records of dairy cows is of particular importance for both breeding and management purposes as they provide a summary of the pattern of milk production with a few parameters that are biologically interpretable in animal science. In this study, hierarchical modeling of the test day milk yield data collected on monthly interval was considered and the non-linear three parameter Wood’s incomplete gamma function was used to obtain the lactation curves of 200 Holstein cows. Bayesian approach was utilized in estimation of model parameters. Some characteristics of lactation such as persistency, peak time and peak yield were also derived as well as total milk yield for each cow.