Home

Services

About Us

Projects

Contact Us

Disclaimer

Project: Dynamic model extraction

The idea behind dynamic model extraction is to use a hybrid approach between statistical regression modeling and detailed mathematical process modeling to determine a system of coupled differential equations whose trajectories span the state space of the data. The objective is to obtain a physically or physiologically meaningful model without the need for detailed process modeling. To use dynamic model extraction a canonical mathematical representation is chosen for the system of differential equations and model parameters are estimated using optimization in the state space over the vector field generated by a data set from a dynamic process. The approach requires longitudinal data sampled over time and does not require that the sampling intervals be equal. Dynamic model extraction methods may be used to determine dynamic models that can produce trajectories consistent with the dynamics of physical or biological processes, even for conditions outside the range of available data. Examples of dynamic model extraction can be found by following the links below.

Douglas-fir forest stand development
Metabolic pathway analysis