Tune#

class abil.tune.tune(X_train, y, model_config, regions=None)#

Bases: object

A class for model training, hyperparameter tuning, and cross-validation.

X_train#

The feature matrix used for training the models.

Type:

pd.DataFrame

y#

The target variable.

Type:

pd.Series

model_config#

Configuration dictionary containing model and training parameters.

Type:

dict

regions#

Name of the feature column representing regions, used for stratification (default is None).

Type:

str or None, optional

train(model, classifier=False, regressor=False, log="no"):

Train and tune models based on the provided configuration.

train(model, classifier=False, regressor=False, log='no')#

Trains a machine learning model using the specified configuration.

Parameters:

  • model (str) – The type of model to train. Supported options: - ‘rf’ : Random Forest - ‘knn’ : K-Nearest Neighbors - ‘xgb’ : XGBoost - ‘gp’ : Gaussian Process

  • classifier (bool, default=False) – Whether to train a classification model.

  • regressor (bool, default=False) – Whether to train a regression model.

  • log (str, default="no") – Log transformation option: - ‘yes’ : Apply log transformation to the target variable. - ‘no’ : No transformation. - ‘both’ : Train both with and without log transformation.

Examples

>>> m.train(model="rf", regressor=True)
Return type:

None