Subscribe Bookmark RSS Feed

JMP 12 : Clustering stacked data does not work?

tajrida

Community Trekker

Joined:

Apr 8, 2015

Hello,

Experimenting on stacked data using the  sample data file "body measurement.JMP" :

1. Used table - stack to stack data in this file.

2. In clustering platform: selected data is stacked, unselected "standardized data" since it is inappropriate for stacked data.

3. continuous data from stacked data column was put in object ID

4. categorical data i.e. label was put in attribute ID

clicked on ok and waited and waited and waited - nothing happens?????

1 ACCEPTED SOLUTION

Accepted Solutions
Solution

In order for this table to work you should create an ID column in the original table before stacking it. This will be used in the Object ID role in the Cluster dialog to identify the individuals to be clustered.

8831_Body_Measurements_jmp_and_untitled_7.png

After stacking you get this:

8832_untitled_7_and_Body_Measurements_jmp.png

Then complete the Cluster dialog like this:

8833_Clustering_and_untitled_7_and_Body_Measurements_jmp.png

The Y Columns identifies the column of measurements to be used to cluster the individuals identified by Object ID. Attribute ID identifies what each row measures.

Your original analysis produced a message in the Log window but should have also resulted in an error dialog. I'll report that and see if we can correct it in a future release.

-Jeff

-Jeff
1 REPLY
Solution

In order for this table to work you should create an ID column in the original table before stacking it. This will be used in the Object ID role in the Cluster dialog to identify the individuals to be clustered.

8831_Body_Measurements_jmp_and_untitled_7.png

After stacking you get this:

8832_untitled_7_and_Body_Measurements_jmp.png

Then complete the Cluster dialog like this:

8833_Clustering_and_untitled_7_and_Body_Measurements_jmp.png

The Y Columns identifies the column of measurements to be used to cluster the individuals identified by Object ID. Attribute ID identifies what each row measures.

Your original analysis produced a message in the Log window but should have also resulted in an error dialog. I'll report that and see if we can correct it in a future release.

-Jeff

-Jeff