Ordinary differential equation (ODE) models in systems biology often comprise unknown parameters which have to be inferred from measurements. This is commonly done by optimizing the agreement of model simulation and measurements. However, experiments often only provide semi-quantitative or even qualitative data. In this thesis, we develop efficient methods for parameter estimation of ODE models based on (i) relative data by combining scalable gradient computation and hierarchical optimization and on (ii) qualitative data by deriving a more efficient framework for the optimal scaling method.
«
Ordinary differential equation (ODE) models in systems biology often comprise unknown parameters which have to be inferred from measurements. This is commonly done by optimizing the agreement of model simulation and measurements. However, experiments often only provide semi-quantitative or even qualitative data. In this thesis, we develop efficient methods for parameter estimation of ODE models based on (i) relative data by combining scalable gradient computation and hierarchical optimization an...
»