BookmarkSubscribeRSS Feed
KenZ
Calcite | Level 5

I am using SAS EM for a logistic regression prediction. I found that sometimes I ran the nodes under the same conditions, but the result can be different (selection of variables ....).

I used stepwise iteration in the logistic regression, and imputed and transformed variables before running the regression node.

Your suggestion will be appreciated!

3 REPLIES 3
JasonXin
SAS Employee

Could you indicate EM version, OS (Windows....), and possibly share the xml diagram file? Jason Xin

Saszealot
Calcite | Level 5

I had similar experience but with SEM 6.1 and 7.1. They generate different outputs.

Sam_zirak
Fluorite | Level 6

If you have partitioned your dataset, you might get different results anytime you run your nodes if the "Rerun" feature on your "Input data" node has been set to "Yes". The reason is that  the partition node splits your dataset into "Training" and "Validation" randomly which could end up in getting slightly different results.

 

Sam

hackathon24-white-horiz.png

The 2025 SAS Hackathon has begun!

It's finally time to hack! Remember to visit the SAS Hacker's Hub regularly for news and updates.

Latest Updates

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
  • 3 replies
  • 1812 views
  • 1 like
  • 4 in conversation