Quantile regression for longitudinal data with a working correlation model

Authors:
Liya Fu;You-Gan Wang
Affiliations:
School of Mathematics and Statistics, Xian Jiaotong University, China and Centre for Applications in Natural Resource Mathematics, School of Mathematics and Physics, The University of Queensland, ...;Centre for Applications in Natural Resource Mathematics, School of Mathematics and Physics, The University of Queensland, St Lucia, QLD 4072, Australia
Venue:
Computational Statistics & Data Analysis
Year:
2012

Citing 3
Cited 2

Quantile regression for longitudinal data

Journal of Multivariate Analysis - Special issue on semiparametric and nonparametric mixed models
Quantile regression without the curse of unsmoothness

Computational Statistics & Data Analysis
Quantile regression for longitudinal data based on latent Markov subject-specific parameters

Statistics and Computing

Smoothing combined estimating equations in quantile regression for longitudinal data

Statistics and Computing
Nonparametric tests for panel count data with unequal observation processes

Computational Statistics & Data Analysis

Quantified Score

Hi-index	0.03

Visualization

Abstract

This paper proposes a linear quantile regression analysis method for longitudinal data that combines the between- and within-subject estimating functions, which incorporates the correlations between repeated measurements. Therefore, the proposed method results in more efficient parameter estimation relative to the estimating functions based on an independence working model. To reduce computational burdens, the induced smoothing method is introduced to obtain parameter estimates and their variances. Under some regularity conditions, the estimators derived by the induced smoothing method are consistent and have asymptotically normal distributions. A number of simulation studies are carried out to evaluate the performance of the proposed method. The results indicate that the efficiency gain for the proposed method is substantial especially when strong within correlations exist. Finally, a dataset from the audiology growth research is used to illustrate the proposed methodology.