Mathematics – Probability
Scientific paper
2004-06-23
Annals of Probability 2004, Vol. 32, No. 3, 2630-2649
Mathematics
Probability
Published at http://dx.doi.org/10.1214/009117904000000153 in the Annals of Probability (http://www.imstat.org/aop/) by the Ins
Scientific paper
10.1214/009117904000000153
Consider a Markov chain on an infinite tree T=(V,E) rooted at \rho. In such a chain, once the initial root state \sigma(\rho) is chosen, each vertex iteratively chooses its state from the one of its parent by an application of a Markov transition rule (and all such applications are independent). Let \mu_j denote the resulting measure for \sigma(\rho)=j. The resulting measure \mu_j is defined on configurations \sigma=(\sigma(x))_{x\in V}\in A^V, where A is some finite set. Let \mu_j^n denote the restriction of \mu to the sigma-algebra generated by the variables \sigma(x), where x is at distance exactly n from \rho. Letting \alpha_n=max_{i,j\in A}d_{TV}(\mu_i^n,\mu_j^n), where d_{TV} denotes total variation distance, we say that the reconstruction problem is solvable if lim inf_{n\to\infty}\alpha_n>0. Reconstruction solvability roughly means that the nth level of the tree contains a nonvanishing amount of information on the root of the tree as n\to\infty. In this paper we study the problem of robust reconstruction. Let \nu be a nondegenerate distribution on A and \epsilon >0. Let \sigma be chosen according to \mu_j^n and \sigma' be obtained from \sigma by letting for each node independently, \sigma(v)=\sigma'(v) with probability 1-\epsilon and \sigma'(v) be an independent sample from \nu otherwise. We denote by \mu_j^n[\nu,\epsilon ] the resulting measure on \sigma'. The measure \mu_j^n[\nu,\epsilon ] is a perturbation of the measure \mu_j^n.
Janson Svante
Mossel Elchanan
No associations
LandOfFree
Robust reconstruction on trees is determined by the second eigenvalue does not yet have a rating. At this time, there are no reviews or comments for this scientific paper.
If you have personal experience with Robust reconstruction on trees is determined by the second eigenvalue, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Robust reconstruction on trees is determined by the second eigenvalue will most certainly appreciate the feedback.
Profile ID: LFWR-SCP-O-247800