Enterprise Miner Decision Tree

[ZaneleD]

Hi

Its my first time building a model. So I am using SAS videos to build the model. I am using one data source . I partitioned it into 60% training, 30% Validation and 10% testing. Now I got an error when running a decision tree.

log Error: Must specify a training or raw data set.;

May you kindly assist: 

Thank you.

 screenshot of the process flow:

[Criptic]

It looks like there is no data being passed into your Decision Tree Node if you click the node and look under Imported Data is it empty? If yes look into your Process and determine why data isn't being passed along (i.e. look into the Exported Data in the previous node).

[ZaneleD]

Thank you for the Reply. The Replacement Node has the imported and exported data. It looks like the data is not passed from the Replacement node to the Control Point node hence it is not in the Decision tree. I clicked Run on the Control Node but still there is no data.

[Criptic]

If you tried to delete the Control Point node and add it back in? I know this sounds dumb but maybe there was some kind of hickup this should work.

[ZaneleD]

Hi

Thank you for the response. i deleted the Control Point node and added it back but still there is an error again. When I click in the import data in the decision tree I can see the data. I cannot find the error in the log. So i do not know on what might be the problem now for the decision tree not to run successful. 

[Criptic]

I'm sorry to say this but without any error message I'm afraid I can not help you. But this behaviour seems highly a typical for SAS if the node is marked as having produced an error there should be an error message in the log.

Have a log at this post: https://communities.sas.com/t5/SAS-Communities-Library/SAS-Enterprise-Miner-shortcut-How-to-find-log... - maybe the error message gets printed to that log

[ZaneleD]

Good Day

It looks like I am experiencing access issues. I will update you once the access issue has been resolved if the decision tree has ran successfully or not.