topic SAS Model Studio - Cluster node in SAS Data Science

SAS Model Studio - Cluster node

prajnat — Wed, 08 Apr 2020 14:32:36 GMT

Hi Team,

Following is the sample data with just 6 columns and 96 rows -

Emp_ID actually needs to be rejected and the variable Q1 - Q5 are inputs

Steps I have tried -

1. To perform any step in the pipelines the model studio is prompting me to define a target variable, in real picture I don't have a target variable. But I tried to create a dummy target variable with no values in it and this threw an error.

2. Just to see the performance of the cluster node, I provided Emp_ID itself as the target and ran the pipeline

3. My output had all the items I was looking for but I am confused as to why the segment plot was performed only on 57 variable while expected segment plot on all 96 rows. Is this due to data partition settings ? if yes, please let me know to edit the window and also confirm if this is a global setting change or the changes I make will be only to specific project

I need to see the segment plot on all the 96 rows.

Please revert

Thank you

Prajna

Re: SAS Model Studio - Cluster node

prajnat — Thu, 09 Apr 2020 05:30:26 GMT

Hi Team,

Can anyone help!

Re: SAS Model Studio - Cluster node

WendyCzika — Thu, 09 Apr 2020 16:54:09 GMT

Yes your data has been partitioned by default, and those results are only on the training observations.

You can see this post for how to re-run your pipeline using different partition settings. You can change your partition settings either in global settings or project settings (it is easiest if you change them for a project before you have run any nodes in a pipeline).

https://communities.sas.com/t5/SAS-Communities-Library/Asked-amp-Answered-How-to-change-data-source-or-data-partition/ta-p/532628