Determining Statistically Significant Parameter Regions for Bounded Specification
- Author(s): Rangarajan, Arvind
- Dornfeld, David A
- et al.
Experiments are performed to observe the influence or quantify the effect a process choice has on the process outcome. Utilization of results from controlled experiments for production plan design using conventional data analysis techniques can lead to inefficient use of available information. The scheme described in this report helps overcome this pitfall by modifying the objective function used to compute the regression parameters. The report also presents a discussion on the various ways of applying this scheme to generate the envelope.