Consistencies and rates of convergence of jump-penalized least squares estimators

Abstract

We study the asymptotics for jump-penalized least squares regression aiming at approximating a regression function by piecewise constant functions. Besides conventional consistency and convergence rates of the estimates in $L^2([0,1))$ our results cover other metrics like Skorokhod metric on the space of càdlàg functions and uniform metrics on $C([0,1])$. We will show that these estimators are in an adaptive sense rate optimal over certain classes of "approximation spaces." Special cases are the class of functions of bounded variation (piecewise) Hölder continuous functions of order $0<\alpha \le 1$ and the class of step functions with a finite but arbitrary number of jumps. In the latter setting, we will also deduce the rates known from change-point analysis for detecting the jumps. Finally, the issue of fully automatic selection of the smoothing parameter is addressed.

Published in at http://dx.doi.org/10.1214/07-AOS558 the Annals of Statistics (http://www.imstat.org/aos/) by the Institute of Mathematical Statistics (http://www.imstat.org)

References in corpus (2)

• Sparse Estimators and the Oracle Property, or the Return of Hodges' Estimator
• Scale space consistency of piecewise constant least squares estimators -- another look at the regressogram

Cited by in corpus (7)

• Wild binary segmentation for multiple change-point detection
• Properties and refinements of the fused lasso
• Nonparametric maximum likelihood approach to multiple change-point problems
• Seeded Binary Segmentation: A general methodology for fast and optimal change point detection
• Smoothing splines for discontinuous signals
• A Variational View on Statistical Multiscale Estimation
• Degrees-of-freedom penalized piecewise regression