I have 6 target classes in my input data and i have specified their data type as nominal.I have tried oversampling and decision node with Gradient boosting, but nothing seems to be producing any result? Are there any specific settings for multi-classification for gradient boosting node in EM?
The boosting node handles nominal targets automatically. No special settings are necessary other than declaring the target as nominal. There should be a result, even if the result is to predict all observations the same way.
But it assigns all classes to 1. and variable importance is empty. I expected GBM results to be better than a single decision tree node.
This is constant prediction. Either no split is created or many trees were created and then pruned back with validation data.
If many trees are created and then pruned, then there should be an iteration sequence.
If no split is created, then yes, some special handling of the settings is necessary. For example, setting the leafsize = 1. Without seeing the data I cannot help much with this.
I am attaching the settings and logs. Let me know, if you need to look into anything else.
Available on demand!
Missed SAS Innovate Las Vegas? Watch all the action for free! View the keynotes, general sessions and 22 breakouts on demand.
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.