|
Activity Number:
|
56
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Sunday, July 29, 2007 : 4:00 PM to 5:50 PM
|
|
Sponsor:
|
IMS
|
| Abstract - #309870 |
|
Title:
|
A Statistical Perspective on Nonlinear Dimensionality Reduction and Manifold Learning
|
|
Author(s):
|
Ann Lee*+
|
|
Companies:
|
Carnegie Mellon University
|
|
Address:
|
Dept of Statistics BH 229J, Pittsburgh, PA, 15213,
|
|
Keywords:
|
manifold learning ; eigenmaps ; kernel PCA
|
|
Abstract:
|
In some problems, the dimension d of the given input space is very large while the data itself has a low intrinsic dimensionality. There are several so called manifold learning methods that aim to learn the geometry of non-linear structures embedded in R^d. Most of these methods, however, approach the data as a fixed set of points and use deterministic techniques to find the optimal embedding coordinates. In this talk, I will present an alternative approach to manifold learning as a statistical estimation problem. Issues such as random noise and regularization will be discussed and put into the context of eigenmaps and kernel PCA.
|
- The address information is for the authors that have a + after their name.
- Authors who are presenting talks have a * after their name.
Back to the full JSM 2007 program |
