Comparison of forecast models of production of dairy cows combining animal and diet parameters

Thong Nguyen (1) , Remy Fouchereau (2) , Emmanuel Frenod (1, 2) , Christine Gerard (3) , Vincent Sincholle (3)

1 LMBA - Laboratoire de Mathématiques de Bretagne Atlantique
2 See-d
3 Neovia

We study the effect of nutritional diet characteristics on the lactating Holstein-Friesian dairy cows in Brittany, France from 36 individuals. An analysis of the relation between fat/protein content and milk yield was conducted. The fat and protein production increase at a slower rate as milk yield increases. The importance of chemical composition and previous observation on milk production forecasting is investigated using both linear models and machine learning approaches (SVM, random forest, neural network). This study evaluates the prediction accuracy of linear model and machine learning based models in both non-autoregressive and autoregressive cases at the individual level. The autoregressive models have proven to have a better performance than the non-autoregressive approaches. Moreover, the computational cost of each approach is presented in the paper. The support vector machine is shown to be the best compromise between accuracy and computational cost.

Mots clés: Milk production forecasting; Dairy modeling; Autoregression; Smart farming

Domaines: Machine Learning [stat.ML]; Applications [stat.AP]; Science des productions animales

Thong Nguyen, Remy Fouchereau, Emmanuel Frenod, Christine Gerard, Vincent Sincholle. Comparison of forecast models of production of dairy cows combining animal and diet parameters. 2019. ⟨hal-02358044v1⟩