Assessment of the future productivity level of dairy cows: between dream and reality

The challenge of increasing the economic and environmental sustainability of the dairy cattle sector involves several factors like milk yield and milk quality levels, cow health and wellbeing, efficient resource use, and emissions reduction. Among the different features, the milk production for lactation, throughout the productive career of a cow, is perhaps the parameter that all farmers would like to know for a more efficient planning of entries and exits from the herd. In fact, if one the one hand numerous researches have studied and are still addressing the problem from a genetic point of view, on the other hand, few studies have focused on the definition of tools for predicting the productivity class future lactations of cow. In the study, firstly two supervised learning methods, i.e., Super Vector Machine and K-Nearest Neighbors, have been applied to a large dataset of 720 complete lactations, with the object to train machine learning tools for the classification between first and second lactation. Then, for those cows having available the data of first and second lactation curve, the two classification methods have been trained and tested for the attribution of the second lactation productivity level (i.e., low, medium or high) starting from the data of the first lactation. The classification methods reached accuracy values ranging from 70% to 73%. These values seem very encouraging and indicate that the predictors selected, despite their simplicity, look very promising and could pave the way for the definition of enhanced future models.