API documentation: flag critical model hyperparameters #20176
Labels
Comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
{{ message }}
Describe the issue linked to the documentation
Some models have many parameters.
Some of those parameters are very important, some others much less so.
While sklearn provides sound defaults usually, it would be nice
if a potential user sees right away which parameters must
be optimized in order to get a significantly better model.
Suggest a potential alternative/fix
For example, in RandomForestsRegressor (and Classifier), n_trees
is a critical parameter (while I have no idea for the other ones, honestly).
In linearSVR, C is a critical parameter (epsilon is not, in my experience).
Let's define a way to tag such parameters in the documentation, and let expert users
tag the critical parameters for models they have experience working with.
The text was updated successfully, but these errors were encountered: