Discussion:
[Scikit-learn-general] yet another parameter for sklearn.tree.DecisionTreeClassifier()
Andreas Mueller
2016-04-05 16:36:05 UTC
Permalink
Hi.
Please post feature discussions on the mailing list or the issue tracker.
For tree based things, Jacob Schreiber, Gilles Louppe and Arnauld Joly
are the experts. You can ping them on github but please don't reach out to
any developers via private email.

(I think post-pruning and stopping using decrease in objective would be
more helpful btw ;)

Andy
Hi Andreas,
I suggest adding one more parameter to
sklearn.tree.DecisionTreeClassifier()
*max_feature_reuse* : int or None, optional (default=None)
The maximum number of times a feature can be reused in the subtree
under the node that splits on this feature.
I think it will be helpful quite often. What do you think?
In case you don't know: whom should I ask?
Cheers,
Boris
Loading...