Online Program Home
My Program

Abstract Details

Activity Number: 103 - New Developments on Statistical Machine Learning
Type: Invited
Date/Time: Monday, July 29, 2019 : 8:30 AM to 10:20 AM
Sponsor: IMS
Abstract #300350 Presentation
Title: Single-Index Thresholding in Quantile Regression
Author(s): Huixia Judy Wang* and Yingying Zhang and Zhongyi Zhu
Companies: The George Washington University and Fudan University and Fudan University
Keywords: Change point; Heterogeneity; Smoothed estimator; Subgroup; Threshold regression

Threshold regression models are useful for identifying subgroups with heterogenous parameters. The conventional threshold regression models split the sample based on a single and observed threshold variable, which enforces the threshold point to be the same for all subgroups of population. To relax this rigid assumption, we consider a more flexible single-index threshold model in the quantile regression setup, where the sample is split based on a single index, an unknown linear combination of predictors. To account for the complication in the asymptotic theory caused by the nonregularity of the model, we propose a smoothed estimator for the regression coefficients and index parameters, and establish its asymptotic properties at a single quantile as well as for the quantile process. Furthermore, we propose a Wald-type test and a mixed-bootstrap procedure for inference on the index parameters.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2019 program