Task 5

This task aims at integrating big data analytics, machine learning, and advanced statistics for the development of a comprehensive analysis model capable of providing real-time support in the optimization of dairy cattle farming and in the design of the relative buildings and plants. Specific objectives are:

• development, test and calibration of a methodology for analysing the big amount of data concerning animal production, behaviour and welfare in relation with environmental conditions;
• development of a forecasting model of animal welfare and production in heat stress conditions;
• the recognition of groups of animals with homogeneous characteristics to optimize the management of the herd;
• early identification of anomalies or problems of production or well-being.

Production, behavior, activity and environmental data are acquired from various dairy farms in the Po Valley having PLF systems (pedometers, collars, rumination monitoring sensors, etc.) and are analysed with advanced big data techniques, machine learning and deep learning. Both dairy farms with milking robot and dairy farms with electronic milking parlour are considered.