Current browse context:
math.PR
Change to browse by:
References & Citations
Mathematics > Probability
Title: Stein's Method for Stationary Distributions of Markov Chains and Application to Ising Models
(Submitted on 15 Dec 2017 (v1), revised 26 Dec 2017 (this version, v2), latest version 14 Sep 2018 (v3))
Abstract: We develop a new technique, based on Stein's method, for comparing two stationary distributions of irreducible Markov Chains whose update rules are `close enough'. We apply this technique to compare Ising models on $d$-regular expander graphs to the Curie-Weiss model (complete graph) in terms of pairwise correlations and more generally $k$th order moments. Concretely, we show that $d$-regular Ramanujan graphs approximate the $k$th order moments of the Curie-Weiss model to within average error $k/\sqrt{d}$ (averaged over the size $k$ subsets). The result applies even in the low-temperature regime; we also derive some simpler approximation results for functionals of Ising models that hold only at high enough temperatures.
Submission history
From: Dheeraj Nagaraj [view email][v1] Fri, 15 Dec 2017 16:40:29 GMT (23kb)
[v2] Tue, 26 Dec 2017 18:35:19 GMT (23kb)
[v3] Fri, 14 Sep 2018 19:22:08 GMT (106kb)
Link back to: arXiv, form interface, contact.