|
Activity Number:
|
192
|
|
Type:
|
Invited
|
|
Date/Time:
|
Tuesday, August 13, 2002 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
ENAR
|
| Abstract - #300118 |
|
Title:
|
Subset Selection in Regression Models using Simulated Annealing: Software and Experiences
|
|
Author(s):
|
Nitin Patel*+
|
|
Affiliation(s):
|
Massachusetts Institute of Technology/Cytel Software Corp.
|
|
Address:
|
675 Massachusetts Avenue, Cambridge, Massachusetts, 02139-3309, USA
|
|
Keywords:
|
data mining ; dimension reduction ; generlized linear regression models
|
|
Abstract:
|
An important dimension reduction technique in data mining is that of selecting subsets of covariates for multiple linear and generalized linear regression models. When the number of covariates to choose from is large, as is often the case with genetic and genomic data, computing best subsets is prohibitively slow. The standard methods used in such situations are variants of step-wise regression. We have adapted the simulated annealing heuristic method to the subset selection problem. Applications to a number of biological data sets has been very encouraging. In this talk we discuss this experience and the easy-to-use software, XL Miner, that we have used in our investigations. Joint work with: Pralay Senchaudhuri (Cytel Software) and Marsha Wilcox (Harvard Medical School).
|
- 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 2002 program |
