-
**context**
In order to achieve full compatibility with JAX we need to make sure that all ensemble gradient estimators actually work with nested parameter containers.
**ideas**
The main problem l…
-
[From documentation]
observation_covariance : [n_dim_obs, n_dim_obs] array-like
Also known as R. observation covariance matrix for times [0...n_timesteps-1]
I'd like to use time-dependent observation…
-
I was thinking it might be a good idea to enrich the positive-definite matrix generation part, by adding correlation matrix generation.
**Interest of covariance matrices**: Covariance and correlati…
-
GEE currently computes all three types of covariance matrices in the model in `_covmat` at the same time.
Instead we could calculate it only based on the `cov_type` that has been chosen by the user i…
-
It seems that both likelihoods in 5_COSMOLOGY/cosmosis_likelihoods use the same covariance matrix (Pantheon+SH0ES_STAT+SYS.cov). I believe that this covariance is the sum of SNIa and Cepheid data cova…
-
Hi again, it'd be nice to be able to run with symmetrical covariance matrices. That is, all gaussians are spheres. I see there is an option to turn off full covariances but this still leaves 3 free pa…
-
Running
`-o Covariance`
_or_
`-o Community -ft Cova -t covariance `
produces matrices indicating they are describing trees, but are I believe, describing bi-partitions.
From the community re…
-
A to-do list of things that still need to be done on the statespace models, to organize future PRs. Anyone is free to add to (or subtract from!) the list.
- [ ] **High priority**: Allow post-estima…
-
allow for multiplications of I to be valid inputted covariance matrices
-
There are many cases where an estimated covariance does not have good small sample properties.
We need some regularization methods (beyond what scikit-learn has)
One of the main tasks will be to prov…