Searching for a trail of evidence in a maze

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 $N(\mu ,1)$, $\mu >0$, and distribution N(0,1) away from it. For which values of the mean shift $\mu$ can one reliably detect and for which values is this impossible? Consider, for example, the usual regular lattice with vertices of the form \[\{(i,j):0\le i,-i\le j\le i and j has the parity of i\}\] and oriented edges $(i,j)\to (i+1,j+s)$, where $s=\pm 1$. We show that for paths of length $m$ starting at the origin, the hypotheses become distinguishable (in a minimax sense) if ${\mu }_m\gg 1/\sqrt{\log m}$, while they are not if ${\mu }_m\ll 1/\log m$. We derive equivalent results in a Bayesian setting where one assumes that all paths are equally likely; there, the asymptotic threshold is ${\mu }_m\approx m^{-1/4}$. 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.

Published in at http://dx.doi.org/10.1214/07-AOS526 the Annals of Statistics (http://www.imstat.org/aos/) by the Institute of Mathematical Statistics (http://www.imstat.org)

References in corpus (4)

• On testing the significance of sets of genes
• Adaptive multiscale detection of filamentary structures in a background of uniform random points
• The Superiority of the Minimal Spanning Tree in Percolation Analyses of Cosmological Datasets
• Detecting Highly Oscillatory Signals by Chirplet Path Pursuit

Cited by in corpus (34)

• Detection of an anomalous cluster in a network
• On combinatorial testing problems
• Detection of correlations
• Cluster detection in networks using percolation
• Rigidity and tolerance for perturbed lattices
• Nonparametric Detection of Geometric Structures over Networks
• Detecting positive correlations in a multivariate sample
• Efficient Minimax Signal Detection on Graphs
• Adaptive sensing performance lower bounds for sparse signal detection and support estimation
• Adaptive Sensing for Estimation of Structured Sparse Signals
• Approximate $l_0$-penalized estimation of piecewise-constant signals on graphs
• Sharp Computational-Statistical Phase Transitions via Oracle Computational Model
• Energy Landscape for large average submatrix detection problems in Gaussian random matrices
• Optimal Detection For Sparse Mixtures
• Quickest Detection of Dynamic Events in Networks
• Separating signal from noise
• Lattice partition recovery with dyadic CART
• Distribution-Free Detection of Structured Anomalies: Permutation and Rank-Based Scans
• Optimal Detection of Random Walks on Graphs: Performance Analysis via Statistical Physics
• Local Two-Sample Testing over Graphs and Point-Clouds by Random-Walk Distributions
• Global testing against sparse alternatives in time-frequency analysis
• Detecting Anomalous Activity on Networks with the Graph Fourier Scan Statistic
• Optimal partition recovery in general graphs
• Detection of Correlations with Adaptive Sensing
• Global Testing Against Sparse Alternatives under Ising Models
• Disagreement-Based Combinatorial Pure Exploration: Sample Complexity Bounds and an Efficient Algorithm
• Clustering Based on Pairwise Distances When the Data is of Mixed Dimensions
• Thresholds For Detecting An Anomalous Path From Noisy Environments
• Sharp Signal Detection Under Ferromagnetic Ising Models
• Detecting the trail of a random walker in a random scenery
• Minimax rates for sparse signal detection under correlation
• On a randomized PNG model with a columnar defect
• Fast and Asymptotically Powerful Detection for Filamentary Objects in Digital Images
• Asymptotic convergence rate of the longest run in an inflating Bernoulli net