|
Activity Number:
|
131
- Simulation and MCMC
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, July 30, 2018 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Statistical Computing
|
|
Abstract #329169
|
|
|
Title:
|
A Direct Quantile Regression
|
|
Author(s):
|
Rachel Morris* and Mei Ling Huang
|
|
Companies:
|
Brock University and Brock University
|
|
Keywords:
|
Conditional quantile;
goodness of fit;
Extreme value distribution;
Gumbel's second kind of bivariate exponential distribution;
nonparametric regression;
loss function
|
|
Abstract:
|
Estimating high conditional quantiles is an important problem. Many studies on this problem use a quantile regression (QR) method. The regular quantile regression method often sets a linear or non-linear model which estimates the coefficients in the model to obtain the estimated conditional quantile. The real-world applications may be restricted by this approach's model setting. This paper proposes a direct nonparametric quantile regression method to overcome this restriction. The paper studies the asymptotic properties of this direct estimator. Monte Carlo simulations show good efficiency for the proposed nonparametric QR estimator relative to the regular QR estimator. The paper also investigates a real-world example of applications by using the proposed method, and gives comparisons of the proposed method and existing methods.
|
Authors who are presenting talks have a * after their name.
