Activity Number:
|
59
- Nonparametric Modeling
|
Type:
|
Contributed
|
Date/Time:
|
Monday, August 3, 2020 : 10:00 AM to 2:00 PM
|
Sponsor:
|
Section on Nonparametric Statistics
|
Abstract #311161
|
|
Title:
|
Quantile Spatially Varying Coefficient Models
|
Author(s):
|
Myungjin Kim* and Lily Wang and Huixia Judy Wang
|
Companies:
|
Iowa State University and Iowa State University and The George Washington University
|
Keywords:
|
Alternating direction method of multipliers;
Bivariate penalized splines;
Nonparametric regression;
Quantile Regression;
Triangulation
|
Abstract:
|
Regression analysis is frequently used in the analyses of spatial data. In this paper, we propose a flexible quantile spatially varying coefficient model to assess how conditional quantiles of the response depend on covariates, allowing the coefficient function varying with the spatial locations. Our model can be used to explore spatial non-stationarity of a regression relationship for heterogeneous spatial data distributed over complex domains. For model estimation, we propose a generic quantile regression method adopting the bivariate penalized spline technique to approximate the unknown functional-coefficients. L2 convergence is established under some regularity conditions and the convergence rate is derived. An efficient algorithm based on the alternating direction method of multipliers (ADMM) is developed to solve the optimization problem. Our Monte Carlo simulation studies have confirmed the excellent performance of the proposed approach. The proposed method is also applied to a mortality dataset to demonstrate performance of the method.
|
Authors who are presenting talks have a * after their name.