Searching for a trail of evidence in a maze
Authors:
Open Access
- 1 August 2008
- journal article
- Published by Institute of Mathematical Statistics in The Annals of Statistics
- Vol. 36 (4) , 1726-1757
Abstract
Consider a graph with a set of vertices and oriented edges connecting pairs of vertices. Each vertex is associated with a random variable and these are assumed to be independent. In this setting, suppose we wish to solve the following hypothesis testing problem: under the null, the random variables have common distribution N(0, 1) while under the alternative, there is an unknown path along which random variables have distribution N(μ, 1), μ> 0, and distribution N(0, 1) away from it. For which values of the mean shift μ can one reliably detect and for which values is this impossible? Consider, for example, the usual regular lattice with vertices of the form {(i, j) : 0≤i, −i≤j≤i and j has the parity of i} and oriented edges (i, j)→(i+1, j+s), where s=±1. We show that for paths of length m starting at the origin, the hypotheses become distinguishable (in a minimax sense) if $\mu_{m}\gg1/\sqrt{\log m}$, while they are not if μm≪1/log m. We derive equivalent results in a Bayesian setting where one assumes that all paths are equally likely; there, the asymptotic threshold is μm≈m−1/4. We obtain corresponding results for trees (where the threshold is of order 1 and independent of the size of the tree), for distributions other than the Gaussian and for other graphs. The concept of the predictability profile, first introduced by Benjamini, Pemantle and Peres, plays a crucial role in our analysis.
Keywords
All Related Versions
This publication has 21 references indexed in Scilit:
- The asymptotics of monotone subsequences of involutionsDuke Mathematical Journal, 2001
- Object tracking using deformable templatesIEEE Transactions on Pattern Analysis and Machine Intelligence, 2000
- Unpredictable nearest neighbor processesThe Annals of Probability, 1998
- Nearest-neighbor walks with low predictability profile and percolation in $2+\epsilon$ dimensionsThe Annals of Probability, 1998
- Unpredictable paths and percolationThe Annals of Probability, 1998
- A spatial scan statisticCommunications in Statistics - Theory and Methods, 1997
- Tree-Indexed ProcessesStatistical Science, 1995
- Directed polymers on trees: a martingale approachJournal of Physics A: General Physics, 1993
- Polymers on disordered trees, spin glasses, and traveling wavesJournal of Statistical Physics, 1988
- Stochastic Comparison of TestsThe Annals of Mathematical Statistics, 1960