ANN is a technology making it possible to build calibrations based on large databases with high complexity. This means that we can without losing any performance, mix different products in one large calibration, which makes our calibration very robust against product changes over time. This allows you to focus on improving your petfood products and controlling your costs, relying on robust and consistent information from a FOSS NIR solution.
The benefits of ANN calibrations compared with traditional PLS calibrations are:
- Inclusion of more information in the same calibration model without loss of accuracy (thousands of samples in the calibration set instead of 100 - 200)
- Ability to solve non-linearity problems
- Ability to handle larger concentration ranges, i.e. fewer calibrations needed to cover a wide range products and raw materials
- A higher degree of transferability between instruments
- Increased robustness and less work in adjusting and up dating the calibration
- Possibility to use calibrations from day one only with minor adjustments
- Accurate and reliable global calibrations.