EFDT debug_one sometimes doesn't work #1247
Replies: 1 comment
-
There's no need to open a discussion an issue and a discussion. I'll close this discussion, and we'll answer in the issue. |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
EDIT: I should have posted this as an issue, I apologize. Reposted here
Hello! I've been experimenting with Extremely Fast Decision Tree.
I have a stream of records and labels. After training my model on each record and label, I use the debug_one method on the previously-labeled records to see if the decision rule had changed. Unfortunately, there is always a problematic record which gives my this error.
It's worth noting that both predict_one and predict_proba_one work. However, the issue seems to affect draw too.
If I train the model a second time on the problematic record, I can use debug_one, but the issue persists with draw.
I don't think the problem lies with the data, because:
Beta Was this translation helpful? Give feedback.
All reactions