|
Activity Number:
|
403
- Sufficient Dimension Reduction and Variable Selection for High-Dimensional Inference
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Wednesday, August 5, 2020 : 1:00 PM to 2:50 PM
|
|
Sponsor:
|
Section on Nonparametric Statistics
|
|
Abstract #309656
|
|
|
Title:
|
Principal Asymmetric Least Squares for Sufficient Dimension Reduction
|
|
Author(s):
|
Yuexiao Dong* and Abdul-Nasah Soale
|
|
Companies:
|
Temple University and Temple University
|
|
Keywords:
|
asymmetric least squares;
expectile regression;
nonlinear dimension reduction
|
|
Abstract:
|
We introduce principal asymmetric least squares (PALS) for linear and nonlinear sufficient dimension reduction. PALS extends the seminal work of principal support vector machines (PSVM) (Artemiou and Li, 2011) and replaces the hinge loss with the asymmetric least squares loss. By synthesizing different expectile levels in the loss function, PALS can handle heteroscedasticity better than PSVM. PALS leads to unbiased estimators at the population level and is attractive computationally. The superior empirical performance of the proposed method is demonstrated through extensive simulation studies and a real data analysis.
|
Authors who are presenting talks have a * after their name.
