|
|
|
# Evidence review
|
|
|
|
|
|
|
|
A page for reviewing the evidence calculation produced by bilby.
|
|
|
|
|
|
|
|
## Analytical likelihood
|
|
|
|
|
|
|
|
To review the evidence calculation using bilby (in particular using the `dynesty` sampler) it can be used to sample from a probability distribution with a known normalisation. In the review of LALInference this consisted of checking the evidence calculation using [two distributions](https://www.lsc-group.phys.uwm.edu/ligovirgo/cbcnote/LALInferenceReviewAnalyticGaussianLikelihood): i) a 15D multivariate Gaussian distribution, and ii) a bi-modal multivariate Gaussian distribution. These same distribution are coded up in bilby in [`15d_gaussian.py`](https://git.ligo.org/lscsoft/bilby/blob/master/examples/core_examples/15d_gaussian.py) and have been tested on [this page](https://git.ligo.org/lscsoft/bilby_pipe/-/wikis/O3-review/15D_Gaussian).
|
|
|
|
|
|
|
|
Review items:
|
|
|
|
|
|
|
|
* run analytical likelihoods multiple times for a range of live points (e.g., 500, 1000, 2000, 5000). Compare empirical uncertainties on log evidence values from the multiple runs to the uncertainty calculated using the K-L divergence (information gain).
|
|
|
|
|
|
|
|
## LALInference comparison
|
|
|
|
|
|
|
|
Compare the evidence produced by the standard bilby run on GW150914 with the LALInference evidence. Run for individual detectors and a coherent multi-detector analysis. |