|
Activity Number:
|
255
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Tuesday, August 4, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Government Statistics
|
| Abstract - #303240 |
|
Title:
|
Using Association Analysis to Enhance Control Chart Applications and Methods
|
|
Author(s):
|
Thomas George*+ and Cathe Reeser
|
|
Companies:
|
Evolver, Inc. and Lockheed Martin
|
|
Address:
|
1943 Isaac Newton Square, Reston , VA, 20190,
|
|
Keywords:
|
Process Control ; Control Charting ; Association Analysis ; Error Identification
|
|
Abstract:
|
Data capture irregularities (both mechanical and operational), can create systematic errors in Census return data. Due to the large quantity of data, the infinite variety of data capture error, and the short time in which to implement continuous improvement methods, detailed analysis of all potential error patterns requires a multi-tiered discovery approach. Broad level response quality can be monitored through control charting. Association analysis can be used to identify anomalies in the infinite number of lower level data patterns. Associations that stray from typical patterns are then highlighted for in-depth analysis and monitored through specifically tailored control charting methods on that particular cross-section until that area of concern is under control.
|
