The QUANCON software
Description.
We propose here implementations of 4 nonparametric methods for estimating conditional quantiles:
- the kernel method, the local constant method and the double kernel method dedicated to unidimensional or multidimensional covariates and non-extreme conditional quantiles,
- a Weissman conditional estimator for functional covariates and extreme conditional quantiles.
The software can be easily interfaced with SAS, Splus or Gnuplot to draw the associated reference curves.
Participants.
- Ali Gannoun (University Montpellier 2).
Publications.
L. Gardes & S. Girard.
"Functional kernel estimators of large conditional quantiles",
Electronic Journal of Statistics, 6, 1715--1744, 2012.
[pdf].
A. Gannoun, S. Girard, C. Guinot & J. Saracco. "Sliced
Inverse Regression
in reference curves estimation", Computational Statistics and Data
Analysis , 46(3):103-122, 2004.
A. Gannoun, S. Girard, C. Guinot & J. Saracco.
"Implémentation en C d'estimateurs non- paramétriques de
quantiles conditionnels. Application au tracé de courbes de
référence", La revue
de Modulad, 31:59-70, 2004.
A. Gannoun, S. Girard, C. Guinot & J. Saracco. "Reference
ranges based
on nonparametric quantile regression", Statistics in Medicine,
21(20):3119-3135, 2002.
A. Gannoun, S. Girard, C. Guinot & J. Saracco. "Trois
méthodes
non paramétriques pour l'estimation de courbes de
référence
- Application l'analyse de propriétés biophysiques de la
peau", Revue de Statistique Appliquée, L(1), 65-89, 2002.
A. Gannoun, S. Girard, C. Guinot & J. Saracco. "Dimension-reduction
in reference curves estimation", Technical report
ENSAM-INRA-University
Montpellier II No 02-02, 2002, [ps].
Download.
- Illustration in R: run.R (functional covariate).
