'parameter estimation - parametric confidence intervals in DAE models
I am using pyomo for parameter estimation where the models are formed by DAEs. It works fine.
Now, I want to construct confidence intervals for parameters, and check the parametric independence. One way is solving the ODEs representing the parametric sensitivity and construct the Fisher Information Matrix. In this case the number of ODEs to solve increases a lot.
Does anybody has experience with a similar estimation and has suggestions? Alternative methodologies are welcome as well as tools that allow compute the sensitivity automatically
Solution 1:[1]
In the meantime I made a few improvements. I read the documentation and parmest.Estimator.theta_est() allows computing the parametric covariance. However it requires pynumero.
I got the error
Cannot load the PyNumero ASL interface (pynumero_ASL)
When I try install it with conda
conda install -c conda-forge/label/cf202003 pynumero_libraries
but the command is not successful. Also tried the suggestions in https://pyomo.readthedocs.io/en/stable/contributed_packages/pynumero/installation.html#method-1 but without success as well.
What can we do? Is there a way to circumvent the need of pynumero? Why am i unable to install it?
Sources
This article follows the attribution requirements of Stack Overflow and is licensed under CC BY-SA 3.0.
Source: Stack Overflow
| Solution | Source |
|---|---|
| Solution 1 | BPMDuarte |