- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
Good day
I am trying the find a way to manually adjust the node rules of a binary classification decision tree using PROC HPSPLIT in SAS EG. More specifically, I am looking to build a model that intuitively and logically splits numerical variables instead of randomly computer generated values i.e. Node 1 split should read variable1 < 200 and variable1 >= 200 as opposed to variable1 < 197.76529 and variable1 >= 197.76529 to create a more easily understood model that can be intuitively implemeted & outputs explained to customers.
I understand this can be done using SAS Enterprise Miner, and SAS EG has the RULES parameter within the PROC HPSPLIT procedure statement that creates an output file containing the node rules applied within the decision tree, however I am uncertain whether the ability exists in SAS EG to do manually adjust these node rules to make better logical sense.
Is anybody aware of how to do this using SAS EG?