|
Activity Number:
|
477
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 9, 2006 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Nonparametric Statistics
|
| Abstract - #307190 |
|
Title:
|
Statistical Inference of Distributions on Manifold
|
|
Author(s):
|
Wanli Min*+
|
|
Companies:
|
IBM T. J. Watson Research Center
|
|
Address:
|
1101 Kitchawan Road, Yorktown Heights, NY, 10598,
|
|
Keywords:
|
non-parametric ; kernel ; dimension reduction
|
|
Abstract:
|
It becomes increasingly common that large dataset of seemingly high dimension can be analyzed through reduced representation on an underlying manifold. This paper considers two related problems, namely, estimation of distribution density from which the random sample are drawn, and construction of the lower dimensional space representation upon which further statistical analysis can be performed . We propose an solution and establish asymptotic properties. An application to digital image analysis will be included.
|
