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

Showing results for

- JMP User Community
- :
- Discussions
- :
- Augmenting an experiment while changing columns de...

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

Dec 8, 2015 6:05 AM
(1592 views)

I have designed a screening experiment assigning some numerical columns to be continuous.

The experiments have been run and I need some supplementary runs however I would like these columns to be discrete numbers.

I am struggling to find a way to include the already existing runs in this new design.

Typically, starting with 4 continuous variables and the 8 runs of a half-fractional experiment, how to generate a design including these runs when 2 of these 4 variables can only take specific values (respectively 3 and 4 levels). The new design is looking at all main effects, secondary interactions, quadratic terms (Xi*Xi) and a blocking var to distinguish the initial runs from the new ones.

1 ACCEPTED SOLUTION

Accepted Solutions

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

Dec 9, 2015 1:20 PM
(2924 views)

Solution

The following should work, but make sure you have your work saved since this is mostly untested. I've tried it in 11.2.1 and 12.1 and have been able to get it to work.

From your original design data table:

1) For the columns you want to be discrete numeric, go to the "design role" and change it to discrete numeric.

2) For those columns, add additional rows with the new levels that you want added. Leave the other columns alone (they'll show up with missing values).

3) Now go to DOE->Augment Design. The factors should be loaded in as discrete numeric with the extra levels you specified in the "fake" rows. These rows are not included in the design because of the missing values.

Hope this helps!

2 REPLIES

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

Dec 9, 2015 1:20 PM
(2925 views)

The following should work, but make sure you have your work saved since this is mostly untested. I've tried it in 11.2.1 and 12.1 and have been able to get it to work.

From your original design data table:

1) For the columns you want to be discrete numeric, go to the "design role" and change it to discrete numeric.

2) For those columns, add additional rows with the new levels that you want added. Leave the other columns alone (they'll show up with missing values).

3) Now go to DOE->Augment Design. The factors should be loaded in as discrete numeric with the extra levels you specified in the "fake" rows. These rows are not included in the design because of the missing values.

Hope this helps!

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

Dec 10, 2015 5:45 AM
(1462 views)