Autocorr and ESS calculation error? about stan HOT 5 CLOSED

Sorry, that should be

src/models/bugs_examples/vol1/magnesium/magnesium --data=src/models/bugs_examples/vol1/magnesium/magnesium.data.R --init=src/models/bugs_examples/vol1/magnesium/magnesium.init.R --seed=489629017; bin/print samples.csv

It seems replicable.

from stan.

I tried this problematic case: it did happen using bin/print; but the
read_stan_csv() function works well.

In addition, bin/print needs either ignore the time information or use
that: now the n_eff/time is inf since the time is 0.

$ print samples.csv
Inference for Stan model: magnesium
1 chains: each with iter=(1000); warmup=(0); thin=(1); 1000 iterations
saved.

Warmup took (0) seconds, 0 seconds total
Sampling took (0) seconds, 0 seconds total

                mean  se_mean       sd  2.5%   25%  50%  75% 97.5%

n_eff n_eff/time Rhat
lp__ -3.0e+03 1.8e+03 3.0e+03-5988.4-5970.7 0.0 0.0 0.0
2.6e+00 inf 10.0
accept_stat__ 2.8e-01 1.7e-01 3.8e-01 0.0 0.0 0.0 0.7 1.0
5.0e+00 inf 1.8
stepsize__ 1.1e-01 6.9e-02 1.1e-01 0.0 0.0 0.0 0.2 0.2
2.6e+00 inf 9.1
treedepth__ 1.5e+00 9.1e-01 1.5e+00 0.0 0.0 0.0 3.0 4.0
2.8e+00 inf 4.9

Jiqiang

On Wed, Jun 5, 2013 at 4:31 PM, Michael Betancourt <[email protected]

wrote:

Sorry, that should be

src/models/bugs_examples/vol1/magnesium/magnesium
--data=src/models/bugs_examples/vol1/magnesium/magnesium.data.R
--init=src/models/bugs_examples/vol1/magnesium/magnesium.init.R
--seed=489629017; bin/print samples.csv

It seems replicable.

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/107#issuecomment-19006404
.

from stan.

Using what's in branch develop now (af5bfad), I don't see this issue with magnesium. Maybe we got it fixed? Can you double check on your machine? If it is fixed, I think it's ok to close the issue.

from stan.

Is this related to the stringstream's not able to parse numbers very very close to 0?

from stan.

@maverickg, possible.

I'm going to close this issue.

from stan.