Stefan Heimersheim
Stefan Heimersheim
Hi everyone! I am not sure if this particular problem has been discussed before but I did do a quick search through the issues and did not find the same....
Description =========== Added a `credibility_interval` method to `MCMCSampels`/`NestedSamples` which computes the credibility / confidence interval. Fixes \#178 and replaces \#179 Notes: ====== - Discretization error: Using `np.cumsum`, the value of...
# Description [This is a draft, mostly for feedback of how to best implement this. I didn't check for typos/functionality, but I have been using these methods myself.] Import dynesty...
**Is your feature request related to a problem? Please describe.** It is not trivial how to compute confidence intervals from `MCMCSamples` or `NestedSamples` even though we have this functionality in...
**Describe the bug** [NestedSamples.importance_sample](https://anesthetic.readthedocs.io/en/latest/anesthetic.html#anesthetic.samples.NestedSamples.importance_sample) changes weights in a weird way. E.g. importance sampling by a constant offset of `logL_new=0.1` changes the weights which it should not (I think) `a=chain.importance_sample(0.1, action="add")`....
Instead of estimating the PDF and drawing contours with plt.contour, we could directly use credible intervals ([fastCI](https://github.com/Stefan-Heimersheim/fastCI)) in y-direction to compute the contours. This avoids typical KDE/PDF-estimation problems like many...
**Describe the bug** I have a sample from my prior (all weights = 1) of a interpolation function (FlexKnot, free-form parameterization etc.) and plot them using fgivenx. (Lists, as I...