No Cover Image

Journal article 1282 views 191 downloads

Estimation of Non-Crossing Quantile Regression Curves

Yuzhi Cai Orcid Logo, Tao Jiang

Australian & New Zealand Journal of Statistics, Volume: 57, Issue: 1, Pages: 139 - 162

Swansea University Author: Yuzhi Cai Orcid Logo

DOI (Published version): 10.1111/anzs.12106

Abstract

Quantile regression methods have been widely used in many research areas in recentyears. However conventional estimation methods for quantile regression models do notguarantee that the estimated quantile curves will be non-crossing. While there are variousmethods in the literature to deal with this...

Full description

Published in: Australian & New Zealand Journal of Statistics
Published: 2015
URI: https://cronfa.swan.ac.uk/Record/cronfa21658
Tags: Add Tag
No Tags, Be the first to tag this record!
Abstract: Quantile regression methods have been widely used in many research areas in recentyears. However conventional estimation methods for quantile regression models do notguarantee that the estimated quantile curves will be non-crossing. While there are variousmethods in the literature to deal with this problem, many of these methods force themodel parameters to lie within a subset of the parameter space in order for the requiredmonotonicity to be satisfied. Note that different methods may use different subspaces of thespace of model parameters. This paper establishes a relationship between the monotonicityof the estimated conditional quantiles and the comonotonicity of the model parameters.We develope a novel quasi-Bayesian method for parameter estimation which can be usedto deal with both time series and independent statistical data. Simulation studies and anapplication to real financial returns show that the proposed method has the potential to bevery useful in practice.
Keywords: asymmetric Laplace distribution; comonotonicity; quasi-Bayesian method
College: Faculty of Humanities and Social Sciences
Issue: 1
Start Page: 139
End Page: 162

Similar Items