Prediction of Milk Yield of Dairy Cows based on Machine Learning

The output of dairy cows will be affected by many factors, it does not only make the management of farm easier, but also assist farmers to can be controlling precisely. The study use history data of each dairy cow from 2018 to 2019, which are 95 dairy cows in the cold season (NTU farm:23、HJ farm:72), 109 in the warm season (NTU farm:25、HJ farm:84), and 84 in the hot season (NTU farm:9、HJ farm:75), some data was discarded due to include missing value. Finally 16121 sets of data were included in the SVR(Support Vector Machine) algorithm prediction model established in this study.

According to dairy cow twice a day can be predicted by the SVR model established by this study with enter the daily activity and environmental heat stress in the past feature parameter, the farmer can identify abnormal status rapidly and make a decision about the nursing work to improve production and the health of dairy cows.

SVR predicted model proposed by the study can provide lots of help on Intelligent farm management in the future.

• Keywords: dairy cows, heat stress, support vector regression(SVR), farm

論文出處連結