JSM 2014 Home
Online Program Home
My Program

Abstract Details

Activity Number: 185
Type: Contributed
Date/Time: Monday, August 4, 2014 : 10:30 AM to 11:15 AM
Sponsor: Section on Statistics in Imaging
Abstract #314010
Title: A Monte Carlo Simulation Study Comparing Robust MSE, Jackknife, and Bootstrapping for Some Robust Regression Estimators in Brain Imaging Research
Author(s): Hung-Wen Yeh*+ and Josh N. Powell and Cary R. Savage
Companies: University of Kansas Medical Center and University of Kansas Medical Center and University of Kansas Medical Center
Keywords: robust regression ; bootstrapping ; jackknife ; brain imaging
Abstract:

Linear regression is widely used to study relationship between brain functions and behavioral or psychometric measures. Acknowledging that ordinary least square estimates is sensitive to outliers, researchers apply robust regression using estimators such as Huber or bi-square to reduce the impact of outliers on regression coefficient estimates. Regression coefficient SE is typically estimated by robust mean square error (MSE). Although literature has shown robust MSE preserves false positive rate (FPR) for univairate outliers (in response variable only), it's not so for bivariate outliers. In this work, we perform simulation for a range of sample sizes in typical fMRI studies to compare computation time and FPR where SE is estimated by robust MSE, jackknife, or bootstrapping. Because brain imaging often involves a large number of voxels and bootstrapping may be time consuming, we investigated different numbers of bootstrap samples to find a balance between computation burden and accuracy. The results provide researchers a guideline in choosing the best method based on their sample size.


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

Back to the full JSM 2014 program




2014 JSM Online Program Home

For information, contact jsm@amstat.org or phone (888) 231-3473.

If you have questions about the Professional Development program, please contact the Education Department.

The views expressed here are those of the individual authors and not necessarily those of the JSM sponsors, their officers, or their staff.

ASA Meetings Department  •  732 North Washington Street, Alexandria, VA 22314  •  (703) 684-1221  •  meetings@amstat.org
Copyright © American Statistical Association.