|
|
# Zero likelihood
|
|
|
|
|
|
To test that bilby samples from the priors, we have a zero likelihood mode where the likelihood only returns 0. A C.I test that this passes is performed every day on the nightly C.I.
|
|
|
To test that bilby samples from the priors, we run a zero-likelihood test where the likelihood only returns 0.
|
|
|
|
|
|
* [The test itself is defined here](https://git.ligo.org/lscsoft/bilby/blob/master/test/sample_from_the_prior_test.py)
|
|
|
* [Click here for the latest results](https://git.ligo.org/lscsoft/bilby/pipeline_schedules). If this is passing then the test is passing |
|
|
\ No newline at end of file |
|
|
The test itself is built into the regular [bilby continuous integration (C.I) tests](https://git.ligo.org/lscsoft/bilby/blob/master/test/sample_from_the_prior_test.py). Broadly speaking it does the following
|
|
|
|
|
|
1. Generates coloured-Gaussian data from the aLIGO design curve
|
|
|
2. Defines a 15D prior for a precessing waveform using `chirp_mass` and `mass_ratio`
|
|
|
3. Generates the normal `bilby` likelihood then replaces it with the zero-likelihood version
|
|
|
4. Runs an inference job (typically takes a few minutes)
|
|
|
5. For each 1D posterior, calculates a p-value comparing the samples in the posterior with 10000 samples drawn from the actual prior.
|
|
|
6. Computes a combined P-value for all 15 individual posteriors
|
|
|
7. If this final p-value is greater than 0.01 the test passes (expected to fail 1/100 runs)
|
|
|
|
|
|
The test is run as part of the Nightly C.I. builds for `bilby. The latest run [can be found here](https://git.ligo.org/lscsoft/bilby/pipeline_schedules). If this is passing then the test is passing |
|
|
\ No newline at end of file |