|
Activity Number:
|
573
|
|
Type:
|
Invited
|
|
Date/Time:
|
Thursday, August 6, 2009 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #302967 |
|
Title:
|
Nonparametric Regression Estimators in Survey Sampling
|
|
Author(s):
|
Sharon Lohr*+ and Guillermo Mendez
|
|
Companies:
|
Arizona State University and American Express
|
|
Address:
|
Department of Mathematics and Statistics, Tempe, AZ, 85287-1804,
|
|
Keywords:
|
classification tree ; random forest ; clustering ; small area estimation
|
|
Abstract:
|
Many survey data sets are large, with many variables as well as complex structure. We explore the use of data mining methods such as regression trees and random forest with survey data. The methods can provide more flexibility for modeling than parametric and local polynomial regression since they are well suited for exploring conditional relationships in the data. We propose inferential methods that account for clustering in the survey data, and show how random forest methods may be used for model-robust small area estimation.
|
