Not checkedNot checked for notices yet.
Prediction of first test day milk yield using historical records in dairy cows
M. Salamone, I. Adriaens, A. Vervaet and 5 more
Animal · 2022
doi:10.1016/j.animal.2022.100658 (opens in a new tab)
We email you when a new notice or a question about this paper appears.
Abstract
The transition between two lactations remains one of the most critical periods during the productive life of dairy cows. In this study, we aimed to develop a model that predicts the milk yield of dairy cows from test day milk yield data collected in the previous lactation. In the past, data routinely collected in the context of herd improvement programmes on dairy farms have been used to provide insights in the health status of animals or for genetic evaluations. Typically, only data from the current lactation is used, comparing expected (i.e., unperturbed) with realised milk yields. This approach cannot be used to monitor the transition period due to the lack of unperturbed milk yields at the start of a lactation. For multiparous cows, an opportunity lies in the use of data from the previous lactation to predict the expected production of the next one. We developed a methodology to predict the first test day milk yield after calving using information from the previous lactation. To this end, three random forest models (nextMILKFULL, nextMILKPH, and nextMILKP) were trained with three different feature sets to forecast the milk yield on the first test day of the next lactation. To evaluate the added value of using a machine-learning approach against simple models based on contemporary animals or production in the previous lactation, we compared the nextMILK models with four benchmark models. The nextMILK models had an RMSE ranging from 6.08 to 6.24 kg of milk. In conclusion, the nextMILK models had a better prediction performance compared to the benchmark models. Application-wise, the proposed methodology could be part of a monitoring tool tailored towards the transition period. Future research should focus on validation of the developed methodology within such tool.
Abstract licence: CC BY-NC-ND
Reads
Read by people on this page
None yet.
Checked or read by AI through naulon
None yet.
A person counts once a day per paper. We keep no cookie and no address: a reader is a hash that changes every day. Any month under five reads shows no number.
Are you an author?
Is this your paper?
Claim it to add your note and be paid when agents cite it. Free.