Distribution-free properties of isotonic regression

  title={Distribution-free properties of isotonic regression},
  author={Jason A. Soloff and Adityanand Guntuboyina and Jim Pitman},
  journal={arXiv: Statistics Theory},
It is well known that the isotonic least squares estimator is characterized as the derivative of the greatest convex minorant of a random walk. Provided the walk has exchangeable increments, we prove that the slopes of the greatest convex minorant are distributed as order statistics of the running averages. This result implies an exact non-asymptotic formula for the squared error risk of least squares in isotonic regression when the true sequence is constant that holds for every exchangeable… Expand

Figures from this paper

Markovian structure in the concave majorant of Brownian motion
The purpose of this paper is to highlight some hidden Markovian structure of the concave majorant of the Brownian motion. Several distributional identities are implied by the joint law of a standardExpand
Adaptation in multivariate log-concave density estimation
We study the adaptation properties of the multivariate log-concave maximum likelihood estimator over two subclasses of log-concave densities. The first consists of densities with polyhedral supportExpand
Abstract. Recently, Chatterjee has introduced a new coefficient of correlation which has several natural properties. In particular, the coefficient attains its maximal value if and only if oneExpand
Correlations with tailored extremal properties
Recently, Chatterjee has introduced a new coefficient of correlation which has several natural properties. In particular, the coefficient attains its maximal value if and only if one variable is aExpand


Risk bounds in isotonic regression
Nonasymptotic risk bounds are provided for maximum likelihood-type isotonic estimators of an unknown nondecreasing regression function, with general average loss at design points. These bounds areExpand
Estimation of a convex function: characterizations and asymptotic theory.
We study nonparametric estimation of convexregression and density functions by methods of least squares (in the regression and density cases) and maximum likelihood (in the density estimationExpand
Sharp oracle inequalities for Least Squares estimators in shape restricted regression
The performance of Least Squares (LS) estimators is studied in isotonic, unimodal and convex regression. Our results have the form of sharp oracle inequalities that account for the modelExpand
Statistical inference under order restrictions : the theory and application of isotonic regression
Abstract : ;Contents: Isotonic regression; Estimation under order restrictions; Testing the equality of ordered means--likelihood ratio tests in the normal case; Testing the equality of orderedExpand
On the risk of convex-constrained least squares estimators under misspecification
We consider the problem of estimating the mean of a noisy vector. When the mean lies in a convex constraint set, the least squares projection of the random vector onto the set is a natural estimator.Expand
Marginal Densities of the Least Concave Majorant of Brownian Motion
A clean, closed form, joint density is derived for Brownian motion, its least concave majorant, and its derivative, all at the same fixed point. Some remarkable conditional and marginal distributionsExpand
Convex minorants of random walks and Lévy processes
This article provides an overview of recent work on descriptions and properties of the Convex minorants of random walks and Levy processes, which summarize and extend the literature on theseExpand
On the quantiles of the Brownian motion and their hitting times
The distribution of the AE-quantile of a Brownian motion on an interval [0, t] has been obtained motivated by a problem in financial mathematics. In this paper we generalize these results by Expand
Asymptotic behavior of the Grenander estimator at density flat regions
Over forty years ago, Grenander derived the MLE of a monotone decreasing density f with known mode. Prakasa Rao obtained the asymptotic distribution of this estimator at a fixed point x where f' (x)Expand
Living on the edge: phase transitions in convex programs with random data
This paper provides the first rigorous analysis that explains why phase transitions are ubiquitous in random convex optimization problems and introduces a summary parameter, called the statistical dimension, that canonically extends the dimension of a linear subspace to the class of convex cones. Expand