Bayesian Network Analysis for the Factors Affecting the 305-day Milk Productivity of Holstein Friesians

Abstract

The variables affecting the milk productivity have been discussed in various articles through different methods. A recent study using path analysis shows that three variables significantly affect the 305-day milk yield of Holstein Friesian cows. These variables are parity, first calving year and lactation length. Calving season is another variable which appears to be significant in a different study. The aim of this study is to provide a simultaneous multilateral analysis among the milk yield, these three variables and a new variable calving season. The analysis was realized through a Bayesian network built over the findings of the path analysis. 17,109 records of Holstein Friesian cows calved between 2001-2011 years were analyzed. The estimated Bayesian network showed that younger cows produced more milk. Lactation length and parity do not depend on each other. Cows reached their highest amount of milk yield on their 4thparities. Milk yield is mostly affected by lactation length. Finally, first calving year, parity, lactation length and calving season should be considered as criteria in a selection study to increase the milk yield.

Keywords

• 305-day milk yield
• holstein friesians
• bayesian networks
• structure learning