BookmarkSubscribeRSS Feed
majeedk
Calcite | Level 5

Below my diagram where I am fitting tree based models. The data node labeled "cfgus_data" is used for training and validation.

My testing data is separate so I added it as node labeled "scoring data" I have also set the output type property to Data and role to test for this data set under "Train" section of properties.

 

What I wanted to get was the "classification table" showing me TP, FP, TN, FN on test data. However, in the "Output" section of results I don't see this for test data set. I am sharing the "Event Classification Table" section below for reference which is only showing the metrics for Train and Validate part.

 

What I am doing wrong? Please advise me how to fix this..

Event Classification Table
Model Selection based on Valid: Average Profit (_VAPROF_)

Model Model Data Target False True False True
Node Description Role Target Label Negative Negative Positive Positive

HPTree HP Tree TRAIN FG 51 455 12 31
HPTree HP Tree VALIDATE FG 20 193 8 16
Tree Decision Tree TRAIN FG 82 467 0 0
Tree Decision Tree VALIDATE FG 36 201 0 0

 

 

Tree.PNG

 

1 REPLY 1
CraigDeVault
SAS Employee

SAS doesn't output the information in the main output window for the test data set.  In order to see the Event Classification table, you can open the Model Comparison node and look at the Fit Statistics table to see the output the test data set.  Another option would be to recreate the counts by running PROC FREQ in a SAS Code node after the Decision Tree node.  The code would look something like this:

proc freq data=&em_import_test;
 tables _NODE_ * I_<targetvar_name>;
run;

Ready to join fellow brilliant minds for the SAS Hackathon?

Build your skills. Make connections. Enjoy creative freedom. Maybe change the world. Registration is now open through August 30th. Visit the SAS Hackathon homepage.

Register today!
How to choose a machine learning algorithm

Use this tutorial as a handy guide to weigh the pros and cons of these commonly used machine learning algorithms.

Find more tutorials on the SAS Users YouTube channel.

Discussion stats
  • 1 reply
  • 2780 views
  • 0 likes
  • 2 in conversation