|
Activity Number:
|
481
- Modeling, Analysis, and Assessment
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 1, 2018 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Quality and Productivity Section
|
|
Abstract #330555
|
Presentation
|
|
Title:
|
A Simulation Study of Process Capability Analysis on Processes with Multiple Normal Distributions
|
|
Author(s):
|
Laura Lancaster*
|
|
Companies:
|
SAS Institute Inc.
|
|
Keywords:
|
Process Capability;
Non-Normal;
Normal Mixture Distributions
|
|
Abstract:
|
Process capability analysis assesses how well a process produces product that is within spec. The standard capability indices relate process performance to spec limits. A process must be in statistical control for these indices to have meaning, and to make inferences about conformance they require the assumption that the measurements come from a normal distribution. It is not uncommon for a process to go out of statistical control due to contamination by a sub-process. Ideally, this contamination could be detected and fixed. However, sometimes the contamination goes undetected or cannot be fixed. The resultant process will have non-normal data. This study will analyze our ability to determine process capability when a process is a mixture of normal distributions. We will compare the performance of two methods for calculating non-normal capability indices that require fitting a distribution as well as the performance of using the standard capability indices that assume normality. We will also study the performance of using a best fit algorithm to determine the distribution before assessing the capability. A variety of normal mixture distributions and sample sizes will be studied.
|
Authors who are presenting talks have a * after their name.
Back to the full JSM 2018 program
|
