On estimation of isotonic piecewise constant signals

DOI10.1214/18-AOS1792MaRDI QIDQ2196185zbMATH OpenOpenAlexFDO

Authors Yanyan Li

Publication date 28 August 2020

Published in The Annals of Statistics (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1705.06386, https://projecteuclid.org/euclid.aos/1590480028

zbMATH Keywords

adaptive estimationisotonic piecewise constant functioniterated logarithmic dependenceoracle inequalitiesreduced isotonic regression

Mathematics Subject Classification ID

Interacting random processes; statistical mechanics type models; percolation theory (60K35) Nonparametric regression and quantile regression (62G08) Asymptotic properties of nonparametric inference (62G20)

Abstract: Consider a sequence of real data points $X_1,ldots,X_n$ with underlying means $het{a}_1^{\ast },dots,het{a}_n^{\ast }$. This paper starts from studying the setting that $het{a}_i^{\ast }$ is both piecewise constant and monotone as a function of the index $i$. For this, we establish the exact minimax rate of estimating such monotone functions, and thus give a non-trivial answer to an open problem in the shape-constrained analysis literature. The minimax rate involves an interesting iterated logarithmic dependence on the dimension, a phenomenon that is revealed through characterizing the interplay between the isotonic shape constraint and model selection complexity. We then develop a penalized least-squares procedure for estimating the vector $het{a}^{\ast }=(het{a}_1^{\ast },dots,het{a}_n^{\ast }{)}^T$. This estimator is shown to achieve the derived minimax rate adaptively. For the proposed estimator, we further allow the model to be misspecified and derive oracle inequalities with the optimal rates, and show there exists a computationally efficient algorithm to compute the exact solution.

Review text

The authors consider observing an $n$-dimensional vector of independent entries with unknown underlying mean. The problem solved is to derive precise minimax rates of the parametric space. They claim that this rate can be achieved by using least-squares procedures, the so-called reduced isotonic regression. They compare it with the ordinary isotonic one and prove that it can avoid overfitting the data. They also derive exact minimax rates under particular losses.

Recommendations

• Isotonic regression in general dimensions
• Estimating piecewise monotone signals
• Isotonic regression in multi-dimensional spaces and graphs
• On risk bounds in isotonic and other shape restricted regression problems
• Nonparametric estimation of piecewise smooth regression functions

Cites work

• A nonparametric changepoint model for stratifying continuous variables under order restrictions and binary outcome
• Adaptation in log-concave density estimation
• Adaptive risk bounds in unimodal regression
• Consistencies and rates of convergence of jump-penalized least squares estimators
• Estimation of unimodal densities without smoothness assumptions
• Exact spike train inference via \(\ell_{0}\) optimization
• FDR-control in multiscale change-point segmentation
• Gaussian model selection
• scientific article; zbMATH DE number 107551 (Why is no real title available?)
• scientific article; zbMATH DE number 193111 (Why is no real title available?)
• scientific article; zbMATH DE number 1220667 (Why is no real title available?)
• scientific article; zbMATH DE number 1048663 (Why is no real title available?)
• scientific article; zbMATH DE number 854949 (Why is no real title available?)
• Introduction to nonparametric estimation
• Least squares estimators of the mode of a unimodal regression function
• Linear time isotonic and unimodal regression in the \(L_{1}\) and \(L_{\infty }\) norms
• Living on the edge: phase transitions in convex programs with random data
• Minimax Rates of Estimation for High-Dimensional Linear Regression Over \ell_q-Balls
• Minimax risk over \(l_ p\)-balls for \(l_ q\)-error
• Near-Optimal Detection of Geometric Objects by Fast Multiscale Methods
• Nonparametric estimation under shape constraints. Estimators, algorithms and asymptotics
• On risk bounds in isotonic and other shape restricted regression problems
• On the degrees of freedom in shape-restricted regression.
• Rates of convergence for minimum contrast estimators
• Risk bounds in isotonic regression
• Sharp oracle bounds for monotone and convex regression through aggregation
• Sharp oracle inequalities for least squares estimators in shape restricted regression
• Sparse estimation by exponential weighting
• The Reduced Monotonic Regression Method
• Unimodal regression using Bernstein–Schoenberg splines and penalties
• Unimodal regression via prefix isotonic regression

Cited in

(19)

• Approximate \(\ell_0\)-penalized estimation of piecewise-constant signals on graphs
• Minimax rates in sparse, high-dimensional change point detection
• On least squares estimation under heteroscedastic and heavy-tailed errors
• Isotonic regression with unknown permutations: statistics, computation and adaptation
• Graph signal denoising using \(t\)-shrinkage priors
• Estimating piecewise monotone signals
• Non-convex isotonic regression via the Myersonian approach
• The bias of isotonic regression
• Multiscale change-point segmentation: beyond step functions
• Isotonic regression in general dimensions
• Estimation of high-dimensional change-points under a group sparsity structure
• Multiscale Quantile Segmentation
• A Mass-Shifting Phenomenon of Truncated Multivariate Normal Priors
• Optimal change-point detection and localization
• Minimax and adaptive tests for detecting abrupt and possibly transitory changes in a Poisson process
• Fused Lasso nearly-isotonic signal approximation in general dimensions
• Empirical priors and posterior concentration in a piecewise polynomial sequence model
• Sharp phase transitions in high-dimensional changepoint detection
• Bayesian uncertainty quantification and structure detection for multiple change points models

Describes a project that uses

Uses Software

• FDRSeg
• isotone
• LZeroSpikeInference

This page was built for publication: On estimation of isotonic piecewise constant signals

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2196185)