I’m trying to compare n_eff computation described in BDA3 and the implemented code, and I think there is a small difference. The comment in the beginning says the implementation matches BDA3. BDA3 (at least version 20150505) describes truncation of the autocorrelation series using Geyer’s initial po…

Ah, found in the manual Exact autocorrelations can happen only on odd lags (Geyer, 2011). By summing over pairs, the paired autocorrelation is guaranteed to be positive modulo estimator noise. This is the motivation behind the many termination criterion of Geyer (2011). Stan does not (yet) do t…

Pull request has been merged. Do RStan, ShinyStan and PyStan use also computation in chains.hpp? Ping @bgoodri @jonah @ariddell

It looks like RStan doesn’t: rstan/rstan/src/chains.cpp. I didn’t realize it wasn’t using Stan’s structure. I think we should consolidate code if we can.

Thanks for your work on chains.hpp, the changes look great. Getting the interfaces to use the same code is or should be a top priority. I tried at one point to have PyStan use chains.hpp but it turned out to be really slow. The main difficulty is exposing some sort function in the services name…

Should I make pull requests for rstan/rstan/src/chains.cpp and pystan/pystan/_chains.pyx ?

Yeah, although it may require more hacking to get it to work.

[image] bgoodri: Yeah, although it may require more hacking to get it to work. Why there is effective_sample_size and effective_sample_size2 ? I can read the comment but I don’t understan it. Is effective_sample_size2 really used?

Hey @avehtari , I am trying to implement this computation in PyMC3. I have a question for you. In your tests, what exactly is dimensions? Is it the length of the chains or number of chains or something else? Thanks in advance! Sharan

[image] sharan: In your tests, what exactly is dimensions? I don’t know what dimensions you are referring. There is not such variable in the parts of the code I modified.

[image] bgoodri: Yeah, although it may require more hacking to get it to work. The more hacking seems to refer to the fact that it’s really difficult to build rstan. There are not much instructions and even @bgoodri keeps guessing what I should do to get it built… PyStan build was easier (on…

If everything you need upstream has been merged into develop, try source("https://raw.githubusercontent.com/stan-dev/rstan/develop/install_StanHeaders.R", echo = TRUE) and then installing the modified rstan.

N_eff BDA3 vs. Stan

Developers

avehtari January 28, 2018, 5:50pm 50

Where does this install those headers?

N_eff greater than number of iterations?

Topic		Replies	Views
How to compare different algorithms for Bayesian inference in terms of speed (and subject to effectiveness) General	16	4077	February 17, 2018
Neff, Rhat for constant looks strange Modeling	23	2639	September 11, 2020
Case study on spatial models for areal data - Poisson CAR/IAR Modeling	116	11618	July 12, 2018
Stan's ESS (and monotone autocorrelations) Algorithms mcmc	5	970	March 13, 2020
Slow Stan Code Modeling fitting-issues , performance	33	3314	June 27, 2018

N_eff BDA3 vs. Stan

Related topics