subsample_high_lnprob

ssapy.utils.subsample_high_lnprob(chain, lnprob, lnprior, nSample, thresh=-10)[source][source]

Select MCMC samples with probabilities above some relative threshold

Parameters:
  • chain (array_like, (nWalker, nStep, 6)) – Input MCMC chain from EmceeSampler.sample()

  • lnprob (array_like, (nWalker, nStep)) – Input MCMC lnprob from EmceeSampler.sample()

  • lnprior (array_like, (nWalker, nStep)) – Input MCMC lnprior from EmceeSampler.sample()

  • nSample (int) – Number of samples to return. If fewer than nSample samples exceed threshold, then return the nSample most probable samples.

  • thresh (float, optional) – Threshold w.r.t. max(lnprob) below which to exclude samples.

Returns:

samples – Output samples

Return type:

array_like, (nSample, 6)