Skip to content

Commit 80efe47

Browse files
authored
Merge pull request #900 from karchjd/patch-1
Fix typos in simulation-based calibration guide
2 parents d1adad0 + e61cde5 commit 80efe47

1 file changed

Lines changed: 2 additions & 2 deletions

File tree

src/stan-users-guide/simulation-based-calibration.qmd

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -10,7 +10,7 @@ contain the true parameter 80% of the time. If data is generated
1010
according to a model, Bayesian posterior inference with respect to
1111
that model is calibrated by construction. Simulation-based
1212
calibration (SBC) exploits this property of Bayesian inference to
13-
asses the soundness of a posterior sampler. Roughly, the way it works
13+
assess the soundness of a posterior sampler. Roughly, the way it works
1414
is by simulating parameters according to the prior, then simulating
1515
data conditioned on the simulated parameters, then testing posterior
1616
calibration of the inference algorithm over independently simulated
@@ -276,7 +276,7 @@ for (k in 1:K) {
276276

277277
The draws from the posterior are assumed to be roughly independent.
278278
If they are not, artifacts may arise in the uniformity tests due to
279-
correlation in the posterior draws. Thus it is best to think the
279+
correlation in the posterior draws. Thus it is best to thin the
280280
posterior draws down to the point where the effective sample size is
281281
roughly the same as the number of thinned draws. This may require
282282
running the code a few times to judge the number of draws required to

0 commit comments

Comments
 (0)