|
Activity Number:
|
449
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 2, 2016 : 2:00 PM to 2:45 PM
|
|
Sponsor:
|
Section on Statistical Learning and Data Science
|
|
Abstract #321684
|
|
|
Title:
|
Modern Projection Pursuit Ellipse for High-Dimensional Data
|
|
Author(s):
|
Jang Ik Cho* and Xiaoyan Wei and Jiayang Sun
|
|
Companies:
|
Case Western Reserve University and Case Western Reserve University and Case Western Reserve University
|
|
Keywords:
|
covariance ;
projection pursuit ;
high-dimensional data ;
large p data
|
|
Abstract:
|
Many standard multivariate techniques were developed based on the assumption that the covariance matrices from different groups are equal. A well-known test for testing the equality of covariance is the Bartlett's test. However the Bartlett's test is only a function of the volumes of covariance matrices not accounting for their shapes and orientations and only for p< < n cases. In this work we developed a Projection Pursuit Ellipses (PPE) for high-dimensional data and compared its performance with the Bartlett's test and a modern benchmark for high dimensional-p data.
|
Authors who are presenting talks have a * after their name.
