search

wayfair / pylift

Uplift modeling package.
http://pylift.readthedocs.io
BSD 2-Clause "Simplified" License
368 stars 76 forks source link

NIV empty #36

Open atersakyan opened 4 years ago

atersakyan commented 4 years ago

Hello,

I've setup my model via TransformedOutcome, then wanted to check all the NIV features. However, when I use NIV (dict, or plot) all the included features are empty. The rest of the steps work and I get plots further down the line, but the NIV being empty leads me to believe they are incorrect.

Setting up the model

up = TransformedOutcome(df, col_treatment='Response', col_outcome='TotalRevenueFoodItems',random_state=4, stratify=y)

Call NIV

up.NIV_dict

Output of NIV: see screenshot

Screen Shot 2019-10-28 at 3 06 23 PM

Thank you.

rsyi commented 4 years ago

Hi atersakayan,

At the moment, NIV is only implemented for binary outcomes.

I'm currently developing on my own branch at https://github.com/pylift/pylift (I'm no longer at wayfair and lost access to the repo), so if you open an issue there I'd be happy to add this functionality in.

atersakyan commented 4 years ago

I don't see a place to open issues on that repo.

Also, does having all zeros in NIV affect the model? Or is that just for EDA?

rsyi commented 4 years ago

Sorry - should be available now.

It doesn't - NIV is just for EDA. :)