Project

Food Safety Prediction

Food safety problems develop due to many factors (called drivers such as climate, economy, human behaviour) acting upon the food chains at different stages and different time points. A system approach is needed to be able to understand these complex systems and to develop methodologies and tools to detect emerging food safety problems at an early stage enabling proper mitigation actions. In a system approach one have to deal with many different data sources having different nature, origin and often large data sets. Besides methodologies that are able to tackle such complexity (i.e. machine learning technologies) and data volumes, also a safe and reliable infrastructure is needed.

Food safety problems develop due to many factors (called drivers such as climate, economy, human behaviour) acting upon the food chains at different stages and different time points. A system approach is needed to be able to understand these complex systems and to develop methodologies and tools to detect emerging food safety problems at an early stage enabling proper mitigation actions. To this end, RIKILT Wageningen UR employs machine leaning methodologies in particularly Bayesian Networks (BN). A BN model will be developed, transferred and tested on a common HPC infrastructure in WUR, including many of the underlying data sources. Where necessary, the common HPC infrastructure will be improved to allow the models to function and to improve the interoperability.

Publicaties