|
Activity Number:
|
418
- From Survival Analysis to Survey Research
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, July 31, 2018 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
International Chinese Statistical Association
|
|
Abstract #326976
|
Presentation
|
|
Title:
|
Pseudo-Kernel Method in Accessing Cross-Validated Risk
|
|
Author(s):
|
Qing Wang*
|
|
Companies:
|
Wellesley College
|
|
Keywords:
|
cross-validation;
model selection;
pseudo-kernel;
U-statistic;
variance estimation
|
|
Abstract:
|
Cross-validation is a widely used tool in accessing the performance of a statistical model or procedure. Most model selection criteria aim at finding the model with the smallest cross-validated risk. Estimating the variance of a cross-validated risk score is crucial when selecting the optimal model in practice. In this paper we devise a pseudo-kernel variance estimator that can be written as a U-statistic based on a pseudo-kernel function of degree two. It is second-order unbiased and easy to implement in the context of K-fold cross-validation. The proposed variance estimator shows comparable performance with significantly improved computational efficiency compared to its bootstrap and jackknife counterparts in simulation and real data analysis in the context of model selection using the "one-standard-error" rule.
|
Authors who are presenting talks have a * after their name.
