This is the program for the 2010 Joint Statistical Meetings in Vancouver, British Columbia.

Abstract Details

Activity Number: 36
Type: Contributed
Date/Time: Sunday, August 1, 2010 : 2:00 PM to 3:50 PM
Sponsor: Section on Nonparametric Statistics
Abstract - #309468
Title: Joint Quantile Regression: A Bayesian Approach
Author(s): Surya Tokdar*+
Companies: Duke University
Address: 426 Euphoria Cir, Cary, NC, 27519, United States
Keywords: Bayesian inference ; Quantile regression ; Nonparametric models ; Logistic Gaussian process ; Tropical cyclones ; Single index model
Abstract:

We consider a joint linear model for all conditional quantile curves of a response variable within a regression setting. For a single predictor the challenging monotonicity constraint is easily met through an interpolation of two monotone curves, modeled via logistic transformations of a Gaussian process. In analyzing time trends of north Atlantic tropical cyclone intensities, we demonstrate our joint model to offer significant improvement over a previously published analysis obtained by fitting separate quantile curves. By borrowing information across the entire conditional distribution, we conclude existence of upward trends of all conditional quantiles -- the previous analysis detected significant trends only in the upper tail. A multivariate extension is proposed and illustrated to offer richer inference than the standard linear heteroskedastic model in analyzing real-world data.


The address information is for the authors that have a + after their name.
Authors who are presenting talks have a * after their name.

Back to the full JSM 2010 program




2010 JSM Online Program Home

For information, contact jsm@amstat.org or phone (888) 231-3473.

If you have questions about the Continuing Education program, please contact the Education Department.