In addition to chemical and technical expertise, the Citizen Science Living Lab also offers knowledge in the field of data science. This knowledge is essential in using the sensor data to create economic or social added value. We can support:
Developing the right data pre-processing - for example of NIR (Near Infra-Red) spectra - and data pre-processing tools.
Developing multivariate (regression) models for determining concentrations of substances. Think of water content, protein content, fat concentration in food or raw materials, etc. We also offer support in determining the correct validation methods, so that the measurements lead to reliable performance figures.
Creating models to classify the measurement results, such as healthy vs. ill or determining the type of food.
How do you process NIR data, so that you can determine reliable concentrations with it? This question seems simple, but in practice it often proves difficult to determine the optimal pre-processing. In collaboration with one of our clients, we have developed a method to make the right choice in types of pre-processing methods - such as baseline correction, scaling, scatter correction - and specific algorithms within them. This saves the customer trial and error experiments and therefore a lot of time.