|
Activity Number:
|
533
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Thursday, August 10, 2006 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #306251 |
|
Title:
|
Automatically Estimating Record Linkage False Match Rates
|
|
Author(s):
|
William E. Winkler*+ and William E. Yancey
|
|
Companies:
|
U.S. Census Bureau and U.S. Census Bureau
|
|
Address:
|
Statistical Research Division, Washington, DC, 20233-9100,
|
|
Keywords:
|
EM algorithm ; unsupervised learning ; semi-supervised learning
|
|
Abstract:
|
This paper provides a mechanism for automatically estimating record linkage false match rates in situations where the subset of the true matches is reasonably well-separated from other pairs. The method provides an alternative to the method of Belin and Rubin (1995) and is applicable in more situations. We provide examples demonstrating why the general problem of error rate estimation (both false match and false nonmatch rates) is likely impossible in situations without training data and exceptionally difficult in the extremely rare situations where training data are available.
|
