parrt / dtreeviz

A python library for decision tree visualization and model interpretation.
MIT License
2.94k stars 331 forks source link

Standardize Capitalization #243

Closed mepland closed 1 year ago

mepland commented 1 year ago

We should standardize the capitalization in the plots, particularly on the axis labels where we have things like feature, leaf ids, samples count as well as GINI. Wouldn't Feature, Leaf IDs, Samples Count, and Gini be preferable?

parrt commented 1 year ago

Agreed. Capitalized per English rules is probably a good idea, unless it's an acronym like MSE. is Gini an acronym? The other caveat is, it should be consistent with the decision tree library. If the library uses 'mse'to indicate the metric then we should use mse not MSE.

mepland commented 1 year ago

Gini is a name, not an acronym. In this case the all uppercase style is coming from the ShadowSKDTree implementation:

    def criterion(self):
        return self.tree_model.criterion.upper()

We should remove the .upper() and take whatever criterion is provided by the model's library.

Sounds good, I'll make a PR to address this issue once https://github.com/parrt/dtreeviz/pull/236 and https://github.com/parrt/dtreeviz/pull/241 are merged!

mepland commented 1 year ago

Addressed in https://github.com/parrt/dtreeviz/pull/245!