turn on suggestions

Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

- JMP User Community
- :
- Discussions
- :
- JMP Mixture Design Parameter Coding

Topic Options

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

Sep 8, 2014 10:47 AM
(671 views)

Basic question: I have an EXTREME VERTEX design for 4 factors generated in JMP. How are the parameters coded? For instance I have a variable (CaCl2) that gets coded in the Parameter table as

(CaCl2-0.3)/0.36

Where do the values "0.3" and "0.36" come from? The mean of the CaCl2 levels is not either number. All the other parameters are coded with "/0.36" in all of them.

Thanks

-Chris

1 ACCEPTED SOLUTION

Accepted Solutions

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

Solution

This is called Lower Pseudocomponent coding. The formula is "Coded CaCl2" = (CaCl2 - the lower bound of CaCl2) / (1 - sum of all lower bounds).

So without the details of this design and your four factors, the lower bound for all of your component ranges add up to 0.64 since 1 - 0.64 = 0.36. The lower bound for CaCl2 is 0.3.

This pseudocomponent coding tries to "expand" the factor space to fill up as much of the design simplex as possible. This will reduce the collinearity of your factors as much as possible, so what JMP is doing is a good thing.

- Dan

Dan Obermiller

2 REPLIES

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

This is called Lower Pseudocomponent coding. The formula is "Coded CaCl2" = (CaCl2 - the lower bound of CaCl2) / (1 - sum of all lower bounds).

So without the details of this design and your four factors, the lower bound for all of your component ranges add up to 0.64 since 1 - 0.64 = 0.36. The lower bound for CaCl2 is 0.3.

This pseudocomponent coding tries to "expand" the factor space to fill up as much of the design simplex as possible. This will reduce the collinearity of your factors as much as possible, so what JMP is doing is a good thing.

- Dan

Dan Obermiller

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

Sep 8, 2014 11:08 AM
(575 views)