Measure concentration for Euclidean distance in the case of dependent random variables
Marton, Katalin
Ann. Probab., Tome 32 (2004) no. 1A, p. 2526-2544 / Harvested from Project Euclid
Let qn be a continuous density function in n-dimensional Euclidean space. We think of qn as the density function of some random sequence Xn with values in $\Bbb{R}^{n}$ . For I⊂[1,n], let XI denote the collection of coordinates Xi, i∈I, and let $\widebar X_{I}$ denote the collection of coordinates Xi, i∉I. We denote by $Q_{I}(x_{I}|\bar{x}_{I})$ the joint conditional density function of XI, given $\widebar X_{I}$ . We prove measure concentration for qn in the case when, for an appropriate class of sets I, (i) the conditional densities $Q_{I}(x_{I}|\bar{x}_{I})$ , as functions of xI, uniformly satisfy a logarithmic Sobolev inequality and (ii) these conditional densities also satisfy a contractivity condition related to Dobrushin and Shlosman’s strong mixing condition.
Publié le : 2004-07-14
Classification:  Measure concentration,  Wasserstein distance,  relative entropy,  Dobrushin–Shlosman mixing condition,  Gibbs sampler,  60K35,  82C22
@article{1091813622,
     author = {Marton, Katalin},
     title = {Measure concentration for Euclidean distance in the case of dependent random variables},
     journal = {Ann. Probab.},
     volume = {32},
     number = {1A},
     year = {2004},
     pages = { 2526-2544},
     language = {en},
     url = {http://dml.mathdoc.fr/item/1091813622}
}
Marton, Katalin. Measure concentration for Euclidean distance in the case of dependent random variables. Ann. Probab., Tome 32 (2004) no. 1A, pp.  2526-2544. http://gdmltest.u-ga.fr/item/1091813622/