Quantile regression via iterative least squares computations

Vito Michele Rosario Muggeo, Mariangela Sciandra, Luigi Augugliaro

Risultato della ricerca: Article

5 Citazioni (Scopus)

Abstract

We present an estimating framework for quantile regression where the usual L1-norm objective function is replaced by its smooth parametric approximation. An exact path-following algorithm is derived, leading to the well-known ‘basic’ solutions interpolating exactly a number of observations equal to the number of parameters being estimated. We discuss briefly possible practical implications of the proposed approach, such as early stopping for large data sets, confidence intervals, and additional topics for future research.
Lingua originaleEnglish
pagine (da-a)1557-1569
Numero di pagine0
RivistaJournal of Statistical Computation and Simulation
Volume82
Stato di pubblicazionePublished - 2012

Fingerprint

Quantile Regression
Least Squares
Early Stopping
Path-following Algorithm
L1-norm
Exact Algorithms
Large Data Sets
Confidence interval
Objective function
Approximation
Least squares
Quantile regression
Framework
Observation

All Science Journal Classification (ASJC) codes

  • Statistics and Probability
  • Applied Mathematics
  • Statistics, Probability and Uncertainty
  • Modelling and Simulation

Cita questo

Quantile regression via iterative least squares computations. / Muggeo, Vito Michele Rosario; Sciandra, Mariangela; Augugliaro, Luigi.

In: Journal of Statistical Computation and Simulation, Vol. 82, 2012, pag. 1557-1569.

Risultato della ricerca: Article

Muggeo, VMR, Sciandra, M & Augugliaro, L 2012, 'Quantile regression via iterative least squares computations', Journal of Statistical Computation and Simulation, vol. 82, pagg. 1557-1569.
Muggeo VMR, Sciandra M, Augugliaro L. Quantile regression via iterative least squares computations. Journal of Statistical Computation and Simulation. 2012;82:1557-1569.
Muggeo, Vito Michele Rosario ; Sciandra, Mariangela ; Augugliaro, Luigi. / Quantile regression via iterative least squares computations. In: Journal of Statistical Computation and Simulation. 2012 ; Vol. 82. pagg. 1557-1569.
@article{a2f989e2fbc84efd9cf9f7b19fea372a,
title = "Quantile regression via iterative least squares computations",
abstract = "We present an estimating framework for quantile regression where the usual L1-norm objective function is replaced by its smooth parametric approximation. An exact path-following algorithm is derived, leading to the well-known ‘basic’ solutions interpolating exactly a number of observations equal to the number of parameters being estimated. We discuss briefly possible practical implications of the proposed approach, such as early stopping for large data sets, confidence intervals, and additional topics for future research.",
keywords = "quantile regression; least squares; smooth approximation",
author = "Muggeo, {Vito Michele Rosario} and Mariangela Sciandra and Luigi Augugliaro",
year = "2012",
language = "English",
volume = "82",
pages = "1557--1569",
journal = "Journal of Statistical Computation and Simulation",
issn = "0094-9655",
publisher = "Taylor and Francis Ltd.",

}

TY - JOUR

T1 - Quantile regression via iterative least squares computations

AU - Muggeo, Vito Michele Rosario

AU - Sciandra, Mariangela

AU - Augugliaro, Luigi

PY - 2012

Y1 - 2012

N2 - We present an estimating framework for quantile regression where the usual L1-norm objective function is replaced by its smooth parametric approximation. An exact path-following algorithm is derived, leading to the well-known ‘basic’ solutions interpolating exactly a number of observations equal to the number of parameters being estimated. We discuss briefly possible practical implications of the proposed approach, such as early stopping for large data sets, confidence intervals, and additional topics for future research.

AB - We present an estimating framework for quantile regression where the usual L1-norm objective function is replaced by its smooth parametric approximation. An exact path-following algorithm is derived, leading to the well-known ‘basic’ solutions interpolating exactly a number of observations equal to the number of parameters being estimated. We discuss briefly possible practical implications of the proposed approach, such as early stopping for large data sets, confidence intervals, and additional topics for future research.

KW - quantile regression; least squares; smooth approximation

UR - http://hdl.handle.net/10447/62348

M3 - Article

VL - 82

SP - 1557

EP - 1569

JO - Journal of Statistical Computation and Simulation

JF - Journal of Statistical Computation and Simulation

SN - 0094-9655

ER -

Accesso al documento