Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

binning_table on a subset of the original sample used for fitting #232

Open
chapmanh opened this issue Feb 17, 2023 · 0 comments
Open

binning_table on a subset of the original sample used for fitting #232

chapmanh opened this issue Feb 17, 2023 · 0 comments
Assignees
Labels
question Further information is requested

Comments

@chapmanh
Copy link

Hi,
This might not be statistically appropriate to do, nonetheless I'm looking to use OptimalBinning to fit a variable to a binary target, and then look at how a subset of the data performs using identical bins to the originally fitted data.

I'm achieving this by providing a new OptimalBinning object the splits obtained in the original fitting and then 'fitting' the subset to these splits with monotonic_trend set to None to prevent shifting. I'm also interested in how the individual IVs change within each bin depending on the subset which is used.

there are three suggestions off the back of this:
one rather radical;

  1. The most radical solution to help me would be some sort of method in binning_table which can ingest a completely different variable array and build a table as if this was the data that created the bins

and two less intense ideas!
2. In instances where the subset does not span the entire range of the original sample, the fitting errors if there are no samples in the bin. Could this default to all zeroes, much like the 'Special' bin already does?
3. Could metric='iv' be an option when using binning_table.plot? This would be really helpful as currently I can only view WoE and event rate within each bin. Thinking about this one, perhaps I just need a crash course on building my own plots from scratch!

Many thanks,
H

@guillermo-navas-palencia guillermo-navas-palencia added the question Further information is requested label Feb 19, 2023
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
question Further information is requested
Projects
None yet
Development

No branches or pull requests

2 participants