A geometrical approach to iterative isotone regression

DOI10.1016/J.AMC.2013.11.048zbMATH Open1364.62094arXiv1211.3930OpenAlexW2018730160MaRDI QIDQ2396476FDOQ2396476

Arnaud Guyader, Alexander B. Németh, S. Z. Németh, Nicolas Jégou

Publication date: 8 June 2017

Published in: Applied Mathematics and Computation (Search for Journal in Brave)

Abstract: In the present paper, we propose and analyze a novel method for estimating a univariate regression function of bounded variation. The underpinning idea is to combine two classical tools in nonparametric statistics, namely isotonic regression and the estimation of additive models. A geometrical interpretation enables us to link this iterative method with Von Neumann's algorithm. Moreover, making a connection with the general property of isotonicity of projection onto convex cones, we derive another equivalent algorithm and go further in the analysis. As iterating the algorithm leads to overfitting, several practical stopping criteria are also presented and discussed.

Full work available at URL: https://arxiv.org/abs/1211.3930

zbMATH Keywords

additive modelsnonparametric estimationisotonic regressionmetric projection onto convex cones

Mathematics Subject Classification ID

Nonparametric regression and quantile regression (62G08)

Cites Work

• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• Estimating the dimension of a model
• Smoothing noisy data with spline functions: Estimating the correct degree of smoothing by the method of generalized cross-validation
• An Empirical Distribution Function for Sampling with Incomplete Information
• Boosting With theL₂Loss
• The asymptotic behavior of monotone regression estimates
• A distribution-free theory of nonparametric regression
• Maximum Likelihood Estimates of Monotone Parameters
• Active set algorithms for isotonic regression; a unifying framework
• Smoothing Parameter Selection in Nonparametric Regression Using an Improved Akaike Information Criterion
• Monotonicity of metric projection onto positive cones of ordered Euclidean spaces
• Linear smoothers and additive models
• An Algorithm for Restricted Least Squares Regression
• Iterative methods for nonlinear complementarity problems on isotone projection cones
• Projection methods, isotone projection cones, and the complementarity problem
• Fitting a bivariate additive model by local polynomial regression
• Dykstra's alternating projection algorithm for two sets
• Additive isotone regression
• On the backfitting algorithm for additive regression models
• How to project onto an isotone projection cone
• On boosting kernel regression
• On the \(\mathbb L_p\)-error of monotonicity constrained estimators
• The Min-Max algorithm and isotonic regression
• An isotonic regression algorithm

Cited In (6)

• Isotonic regression and isotonic projection
• Extended Lorentz cones and variational inequalities on cylinders
• A geometric interpretation of Darroch and Ratcliff's generalized iterative scaling
• Iterative isotonic regression
• Extended Lorentz cones and mixed complementarity problems
• Projection onto simplicial cones by Picard's method

Uses Software

• ibr

This page was built for publication: A geometrical approach to iterative isotone regression

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2396476)