Gelman-Rubin statistic
{{Expert needed|1=Mathematics|date=January 2024}}
The Gelman-Rubin statistic allows a statement about the convergence of Monte Carlo simulations.
== Definition ==
Monte Carlo simulations (chains) are started with different initial values. The samples from the respective burn-in phases are discarded.
From the samples (of the j-th simulation), the variance between the chains and the variance in the chains is estimated:
: Mean value of chain j
: Mean of the means of all chains
: Variance of the means of the chains
: Averaged variances of the individual chains across all chains
An estimate of the Gelman-Rubin statistic then results as{{Cite book|url=https://bookdown.org/rdpeng/advstatcomp/monitoring-convergence.html|title=7.4 Monitoring Convergence | Advanced Statistical Computing|first=Roger D.|last=Peng|via=bookdown.org}}
:.
When L tends to infinity and B tends to zero, R tends to 1.
A different formula is given by Vats & Knudson.{{cite journal | doi = 10.1214/20-STS812| title = Revisiting the Gelman–Rubin Diagnostic| date = 2021| last1 = Vats| first1 = Dootika| last2 = Knudson| first2 = Christina| journal = Statistical Science| volume = 36| issue = 4| arxiv = 1812.09384}}
== Alternatives ==
The Geweke Diagnostic compares whether the mean of the first x percent of a chain and the mean of the last y percent of a chain match.{{citation needed|date= January 2024}}
Literature
- {{cite journal |doi=10.1214/20-STS812 |title=Revisiting the Gelman–Rubin Diagnostic |date=2021 |last1=Vats |first1=Dootika |last2=Knudson |first2=Christina |journal=Statistical Science |volume=36 |issue=4 |arxiv=1812.09384 }}
- {{cite journal |doi=10.1214/ss/1177011136 |title=Inference from Iterative Simulation Using Multiple Sequences |date=1992 |last1=Gelman |first1=Andrew |last2=Rubin |first2=Donald B. |journal=Statistical Science |volume=7 |issue=4 |pages=457–472 |jstor=2246093 |bibcode=1992StaSc...7..457G }}
== References ==