logistic_regression_ovr

besca.tl.auto_annot.logistic_regression_ovr(train, y_train, njobs)

multiclass crossvalidated logistic regression with balanced class weight. Requires ten nodes. Returns OVR probability.

Parameters:

• train (pd.DataFrame) – adata_train.X but scaled and as a dataframe

• y_train (pd.DataFrame) – one dimensional dataframe containing class label

Returns:

trained

Return type:

sklearn.calibration.CalibratedClassifierCV