[BUG] Can't set positive_class for tabular predictor #5114
Description
Bug Report Checklist
- I provided code that demonstrates a minimal reproducible example.
- I confirmed bug exists on the latest mainline of AutoGluon via source install.
- I confirmed bug exists on the latest stable version of AutoGluon.
Describe the bug
Users can't set the positive_class for the TabularPredictor.
positive_class isn't in the set of kwargs since it is directly specified in the init. You can probably update
autogluon/tabular/src/autogluon/tabular/predictor/predictor.py
Lines 209 to 210 in 73173f6
To just:
learner_kwargs["positive_class"] = positive_class
or pass it directly to the AbstractTabularLearner type in the downstream call
Expected behavior
User's should be able to set the positive_class
To Reproduce
from autogluon.tabular import TabularDataset, TabularPredictor
data_root = 'https://autogluon.s3.amazonaws.com/datasets/Inc/'
train_data = TabularDataset(data_root + 'train.csv')
test_data = TabularDataset(data_root + 'test.csv')
predictor = TabularPredictor(label='class', positive_class='<=50K').fit(train_data=train_data)
predictions = predictor.predict(test_data)
Screenshots / Logs
in TabularPredictor.__init__(self, label, problem_type, eval_metric, path, verbosity, log_to_file, log_file_path, sample_weight, weight_evaluation, groups, positive_class, **kwargs)
208 quantile_levels = kwargs.get("quantile_levels", None)
209 if positive_class is not None:
--> 210 learner_kwargs["positive_class"] = kwargs["positive_class"]
212 self._learner: AbstractTabularLearner = learner_type(
213 path_context=path,
214 label=label,
(...) 222 **learner_kwargs,
223 )
224 self._learner_type = type(self._learner)
KeyError: 'positive_class'
Installed Versions
# Replace this code with the output of the following:
from autogluon.core.utils import show_versions
show_versions()