EvgeniDubov / hellinger-distance-criterion

Random Forest model using Hellinger Distance as split criterion
BSD 3-Clause "New" or "Revised" License
31 stars 12 forks source link

Update hellinger_distance_criterion.pyx #13

Closed 0zean closed 1 month ago

0zean commented 1 year ago

Updated to work with latest Scikit-learn _criterion.pyx version 1.x and Cython 3.0+ (implicit noexcept declaration is deprecated)

harish1996 commented 1 year ago

Kindly take a look at this issue https://github.com/EvgeniDubov/hellinger-distance-criterion/issues/4. I am yet to take a look at scikit-learn 1.x, but are you sure Cython supports a list as a subscript ?

harish1996 commented 1 year ago

I just went through the changes in scikit-learn and cython 3+, and it looks like i rushed while commenting previously. The fixes suggested in this looks to support scikit-learn > 1.1.0.

Relevant PR

I havent tested it. Will try to test it soon.

Although i wasnt able to find any mention of this change in any of the changelogs! Is tweaking criterion outside of sklearn undocumented ?