Guillaume Lemaitre
Guillaume Lemaitre
This PR proposes the following refactoring: - take profit from the helper `_get_response_values` instead of handling manually the `response_method`. - remove a double check to know if the estimator provided...
This is a bit hackish but this the way to keep the title with the version number and show the date on a second line in bold as in the...
This is updating our documentation after introducing `towncrier` into our PR process. TODO: - [x] update the saved replies to add entry in the changelog - [ ] update the...
closes #26482 closes #26691 superseded #26691 While reviewing #26691, we saw that there is a coupling between the parameter `dictionary` in `SparseCoder` and `X`. Therefore, the simplest way to not...
While reviewing a test in https://github.com/scikit-learn/scikit-learn/pull/16937, it appears that our implementation of `_weighted_percentile` with unit `sample_weight` will lead to a different result than `np.median` which is a bit problematic for...
closes #30027 When no early-stopping is used in SGD estimator, the stopping criterion should be based on the value of the objective function (loss function + regularization terms). However, currently...
This is a follow-up to the issue raised in https://github.com/scikit-learn/scikit-learn/issues/29554. However, I recall other issues raised for CV estimator in general. So the context is the following: a CV estimator...
This PR add the necessary class and function to add a link to the documentation when showing the HTML representation of an estimator. Unfortunately, there is a bug in scikit-learn...
When using `numpydoc_show_class_members = True` in the config, the class methods are listed in a table as follow:  In scikit-learn, we recently switch to `pydata-sphinx-theme` that has a right-hand...
In #30763, we reorganize the module for the `repr` of estimator. The public functions associated with this module and that could be potentially used by developer of third-party libraries should...