13th European Conference on Mathematical and Theoretical Biology - ECMTB 2024, Toledo (Spain). 22-26 July 2024
Summary:
Mathematical modelling is a valuable tool in studying viral dynamics, allowing us to comprehend complex mechanisms at various levels and make predictions about their behaviour. However, constructing an accurate model that represents a system can pose challenges. One crucial aspect to consider is the level of detail necessary for the model to be interpretable because, in many cases, adding more layers of complexity renders the model unidentifiable, meaning it cannot be uniquely estimated from the available data. In this presentation, we will explore two significant issues regarding identifiability: (i) the effective information contained within the data and how this limitation impacts the modelling process, and (ii) the consequences of information dissipation resulting from data aggregation. Several case studies will be examined to illustrate the challenges and opportunities in finding the right balance between model complexity and identifiability. Additionally, we will point out examples where simplicity outweighs excessive focus on details.
Publication date: 2024-07-22.
Citation:
M. Castro, Data-dependent limits to modeling viral dynamics and epidemic spreading, 13th European Conference on Mathematical and Theoretical Biology - ECMTB 2024, Toledo (Spain). 22-26 July 2024.
