27/04/2026
Our next VistaMilk impact case study highlights how advanced statistical modelling is improving our ability to predict milk processing performance.
This work focuses on the development of a Bayesian probabilistic least squares regression (bplsr) approach to predict milk heat stability — a key factor in determining how milk performs during processing and its suitability for different products.
By reducing prediction error and better capturing the complexity of milk composition data, this model provides a more reliable tool for industry decision-making, supporting improved product consistency and processing efficiency.
A key strength of this work is its accessibility. The model has been developed as an open-source tool, enabling uptake beyond the research environment and supporting wider application across the dairy sector.
It is a strong example of how combining data science with dairy research can deliver practical, industry-relevant outcomes.