Science in Three Minutes: Smarter and Healthier Dairy Farming with Machine Learning

Calf diseases are often detected only when they have already progressed significantly. Late diagnosis leads to increased antibiotic use, slower growth, and economic losses. At the same time, traditional methods such as measuring body temperature in large barns are labor-intensive and stressful for the animals.

Rait Rand’s research focuses on the use of subcutaneous sensors combined with machine learning systems that enable earlier disease detection. Machine learning models can distinguish environmental influences from an animal’s actual physiological condition, allowing body temperature to be assessed with clinical accuracy without invasive methods.

In addition, a metric called “fever load” is used, which evaluates not only the rise in body temperature but also its duration and intensity. This helps distinguish random fluctuations from actual signs of disease and makes it possible to detect illness before visible symptoms appear.

Rait Rand’s doctoral research explores how smart technologies and a data-driven approach can transform livestock farming. Earlier disease detection enables more precise treatment, reduced antibiotic use, and more efficient farm management. The doctoral thesis is supervised by Associate Professor of Population Medicine Tarmo Niine.

Watch the video in which doctoral student Rait Rand explains how machine learning helps improve herd health monitoring. His presentation was awarded third place at the Estonian University of Life Sciences “Science in Three Minutes” competition (switch on ENG subtitles in YouTube):