— Back to Blogs & Publications Evaluating Cheese Yield Prediction Models Using caliX

Evaluating Cheese Yield Prediction Models Using caliX

Maximize cheese manufacturing yield by predicting vat outcomes using chemometric modelling.

Why Cheese Yield Prediction Matters

Small variations in milk composition (fat, casein, lactose) and coagulation parameters lead to significant differences in final cheese yield. Over-retaining moisture leads to soft defects; over-pressing loses fat in the whey drainage. Predicting final curd yield before pressing allows cheese makers to adjust cutting, heating, and salting times in the vat.

Building PLS Yield Models in caliX

By compiling spectroscopic scans of vat milk and corresponding lab results using the caliX Spectral Intelligence suite, the dairy plant constructed a robust PLS regression model. The model predicts moisture-in-curd with high accuracy, enabling automated decisions on vat draw-down timings.

Parameter Range Root Mean Square Error of Cross-Validation (RMSECV)
Curd Moisture 38.0% — 46.0% —0.18%
Fat-in-Dry-Matter (FDM) 48.0% — 54.0% —0.25%

References

  • "Multivariate Chemometric Calibrations for Coagulation Kinetics in Cheddar Cheesemaking," Dairy Research Communications, 2024.
USTECH Innovations Logo

Download Full Technical Report

Access the complete, unedited technical paper with full spectroscopic formulas, calibration datasets, and instrumentation parameters.

Read Next