pyomo
pyomo copied to clipboard
Pyomo
Using parmest to estimate experiment specific parameters
Summary
Are there examples of using parmest to simultaneously estimate both "global" parameters across all experiments (stage 1 in the stochastic programming abstraction, i.e., what is done by default when specifying theta) and individual/local parameters for each experiment (stage 2 in the stochastic programming abstraction)?
Rationale
This question is motivated by a reaction kinetics project in which we want to estimate a "global" set of kinetic parameters while also fitting initial conditions for each experiment (with modest bounds).
Description
Were are able to do this by passing the stage 1 parameter values to parmest as theta and leaving the stage 2 parameters as free variables (with bounds). While this works, it does not properly compute the covariance matrix.
Additional information
Links to any examples are greatly appreciated. Otherwise, I'd like to discuss here what would be involved in adding this as a feature.
