search

pybop-team / PyBOP

A parameterisation and optimisation package for battery models.
https://pybop-docs.readthedocs.io
BSD 3-Clause "New" or "Revised" License
66 stars 20 forks source link

Point parameter estimation methods #5

Closed BradyPlanden closed 8 months ago

BradyPlanden commented 1 year ago

Overarching parent issue to cover the family of all methods that give only "point" parameter estimates (rather than distributions) - i.e., maximum likelihood estimation (MLE) including variations such as prediction error minimisation (PEM); maximum a posteriori (MAP) estimation.

BradyPlanden commented 1 year ago

This is a parent issue for #3.

martinjrobins commented 1 year ago

in terms of a method to get the point estimate, CMAES is a pretty reliable method (again lots of implementations including PINTS ;)

davidhowey commented 1 year ago

interesting; we should have a deep dive discussion on all this. We have tended previously just to use variants of BFGS

BradyPlanden commented 8 months ago

We have added quite a few frequentistic methods, closing this issue.