great idea,
they really need to do this
meantime you can use
https://imbalanced-learn.org/stable/auto_examples/applications/plot_impact_imbalanced_classes.html

sample weights support will be available in 0.23 which should be out soon.
So you can have class weights through that.

Just a reminder

they really need to do this

Let's avoid that kind of phrasing please.

great news thanks
by the way will this coming solution to have the same problem what lightgbm and xgboost have?
microsoft/LightGBM#2107
scale_pos_weight is used to adjust the imbalance will spoil probability?

as it is written : result in poor estimates of the individual class probabilities
https://lightgbm.readthedocs.io/en/latest/Parameters.html

is_unbalance 🔗︎, default = false, type = bool, aliases: unbalance, unbalanced_sets

used only in binary and multiclassova applications

set this to true if training data are unbalanced

Note: while enabling this should increase the overall performance metric of your model, it will also result in poor estimates of the individual class probabilities

or
consider performing probability calibration (https://scikit-learn.org/stable/modules/calibration.html)

https://lightgbm.readthedocs.io/en/latest/pythonapi/lightgbm.LGBMClassifier.html
class_weight (dict, 'balanced' or None, optional (default=None)) – Weights associated with classes in the form {class_label: weight}. Use this parameter only for multi-class classification task; for binary classification task you may use is_unbalance or scale_pos_weight parameters. Note, that the usage of all these parameters will result in poor estimates of the individual class probabilities. You may want to consider performing probability calibration (https://scikit-learn.org/stable/modules/calibration.html) of your model. The ‘balanced’ mode uses the values of y to automatically adjust weights inversely proportional to class frequencies in the input data as n_samples / (n_classes * np.bincount(y)). If None, all classes are supposed to have weight one. Note, that these weights will be multiplied with sample_weight (passed through the fit method) if sample_weight is specified

Came here to look for this Issue and add my +1!

@NicolasHug wrote:

sample weights support will be available in 0.23 which should be out soon. So you can have class weights through that.

Happy to have sample_weights support, but that doesn't really help for my use case. Class weights generally need to be tuned like a hyperparameter, and I can't do that using, say, GridSearchCV because there is no class_weight param.