Skip to content
Snippets Groups Projects

Add a mean-log-likelihood method to improve the ACT estimation

Merged Gregory Ashton requested to merge add-mean-log-like-to-ptemcee into master
All threads resolved!
Compare and Show latest version
1 file
+ 11
5
Compare changes
  • Side-by-side
  • Inline
@@ -90,6 +90,12 @@ class Ptemcee(MCMCSampler):
The walkers are then initialized from the range of values obtained.
If a list, for the keys in the list the optimization step is applied,
otherwise the initial points are drawn from the prior.
niterations_per_check: int (5)
The number of iteration steps to take before checking ACT. This
effectively pre-thins the chains. Larger values reduce the per-eval
timing due to improved efficiency. But, if it is made too large the
pre-thinning may be overly agressive effectively wasting compute-time.
If you see tau=1, then niterations_per_check is likely too large.
Other Parameters
@@ -98,7 +104,7 @@ class Ptemcee(MCMCSampler):
The number of walkers
nsteps: int, (100)
The number of steps to take
ntemps: int (2)
ntemps: int (10)
The number of temperatures used by ptemcee
Tmax: float
The maximum temperature
@@ -107,7 +113,7 @@ class Ptemcee(MCMCSampler):
# Arguments used by ptemcee
default_kwargs = dict(
ntemps=20,
ntemps=10,
nwalkers=200,
Tmax=None,
betas=None,
@@ -130,8 +136,8 @@ class Ptemcee(MCMCSampler):
skip_import_verification=False,
resume=True,
nsamples=5000,
burn_in_nact=10,
thin_by_nact=0.5,
burn_in_nact=50,
thin_by_nact=1,
autocorr_tol=50,
autocorr_c=5,
safety=1,
@@ -145,7 +151,7 @@ class Ptemcee(MCMCSampler):
store_walkers=False,
ignore_keys_for_tau=None,
pos0="prior",
niterations_per_check=10,
niterations_per_check=5,
**kwargs
):
super(Ptemcee, self).__init__(
Loading