|
Activity Number:
|
544
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 10, 2006 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Statistical Graphics
|
| Abstract - #306478 |
|
Title:
|
Design Strategies for Sampling in Graphs
|
|
Author(s):
|
James Rosenberger*+ and Hong Xu and Steve Thompson
|
|
Companies:
|
The Pennsylvania State University and Simon Fraser University
|
|
Address:
|
326 Thomas Building, University Park, PA, 16802-2111,
|
|
Keywords:
|
networks
|
|
Abstract:
|
Sampling graphs is an important challenge not adequately addressed in the literature. Inference about graphs often confuses the sample data with the entire population. The data graph is usually a single realization, not representative of the population graph as a whole. Designs for sampling graphs are important for efficient estimation. Thompson (2004) constructed the Active Set Adaptive Sampling (ASAS) designs and associated estimators, which can apply to populations with network structures. The main advantage is that the sample size can be predefined and fixed prior to implementing the sampling strategy. In this paper, we incorporate some practical restrictions, such as budget or time constraints that determine which units should be selected at each step and when to stop the sampling process. We want to minimize the cost and maximize the sampling procedure information.
|
