With recent technological advancement, mathematical models are being widely applied to support decision-making in the life sciences, e.g. for planning experiments or intervention strategies for disease spreading. As such it is increasingly important that sources of errors are investigated and reported in these models, and users (both expert and non) are equipped to critically assess them. There is no well known standard available for mathematical model assessment and comparison in application, and the vision of this proposed research is to offer guidelines for good practice and application-oriented standards for model evaluation for non-expert users that take the specificities of the modeling and analysis pipeline into account.
We aim to identify starting points for this task through the exploration of uncertainty throughout the modeling pipeline leading to the identification of key sources of uncertainty such as model topology, parameter identification and data quality that can be generalized to different application areas drawing on the collaborators’ research expertise.