Skip to content

Conversation

@GregoryAshton
Copy link
Collaborator

@GregoryAshton GregoryAshton commented Apr 3, 2025

  • Adds a sampler wrapper so that the fisher methods can be called using run_sampler for ease of use.
  • Adds a rejection sample method to reweight to the posterior
  • Adds diagnostic plots
  • Improve the maximum likelihood estimation method
  • Replace the naive finite difference Hessian estimator with scipy.differentiate function

@GregoryAshton GregoryAshton changed the title Add a fisher-information-matrix sampler Draft: Add a fisher-information-matrix sampler Apr 3, 2025
@GregoryAshton
Copy link
Collaborator Author

Note, this is the corner plot created:

image

- Add use of scipy.differentiate for Hessian computation
- Fix bugs in sampler implementation
…ters

If a pandas dataframe with a non-monotonically increasing index (say if
it has been resampled) is passed in, the new_samples data frame that is
created will have a different index and when matched NaN's will be
introduced. This fixes that issue by setting the column from the values
Due to the division of 0/0, this produced infs in the calculate tilt
parameters.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant