Recommender System for Healthy Food Choices

A person's arm with and phone with the application

A recommender system that proposed the right meal at the right time based on your biomarkers and contextual information? That is the challenge that this research project is tackling together with Clear. Clear is a startup in nutrition precision and has a unique dataset of user demographics, logged meals and postprandial glucose values of these meals. The data confirms what is known from recent literature: people can respond differently to the same meals.

But this leaves many questions still unanswered: 

  • Can we predict the postprandial glucose values based on previous glucose measurement, user demographics and meal composition alone?
  • Can we identify similar responses to meals and use this to cluster users and create more accurate predictions? 
  • Can we translate this prediction model into a recipe recommender system that takes into account food preferences, allergies and contextual information?

Researchers

Hanna Hauptmann, Christine Bauer

Partner

Clear B.V.