actableai.models.autogluon.params package

Submodules

actableai.models.autogluon.params.ag_automm module

class actableai.models.autogluon.params.ag_automm.AGAUTOMMParams

Bases: actableai.models.autogluon.params.base.BaseParams

Parameter class for AG_AUTOMM Model.

explain_samples_supported: bool = False

gpu_required: bool = True

supported_problem_types: Literal['regression', 'binary', 'multiclass', 'quantile'] = ['regression', 'binary', 'multiclass']

actableai.models.autogluon.params.base module

class actableai.models.autogluon.params.base.BaseParams

Bases: abc.ABC

Base class for Regression Model Parameters.

explain_samples_supported: bool

classmethod get_autogluon_name() → Union[str, Type[AbstractModel]]

classmethod get_autogluon_parameters(hyperparameters: Dict[str, Any], model_hyperparameters_space: actableai.parameters.parameters.Parameters, process_hyperparameters: bool = True) → dict

Converts hyperparameters for use in AutoGluon’s hyperparameter optimization search.

Parameters

• params – The hyperparameters to be used by the model
• model_hyperparameters_space – The model’s hyperparameter space
• process_hyperparameters – If True the hyperparameters will be validated and processed (deactivate if they have already been validated).

Returns

dictionary with hyperparameters in AutoGluon

format.

Return type

parameters_autogluon

classmethod get_hyperparameters(*, problem_type: str, device: str, num_class: int) → actableai.parameters.parameters.Parameters

Returns the hyperparameters space of the model.

Parameters

• problem_type – Defines the type of the problem (e.g. regression, binary classification, etc.).
• device – Which device is being used, can be one of ‘cpu’ or ‘gpu’
• num_class – The number of classes, used for multi-class classification.

Returns

The hyperparameters space.

gpu_required: bool = False

supported_problem_types: Literal['regression', 'binary', 'multiclass', 'quantile']

class actableai.models.autogluon.params.base.Model(value)

Bases: str, enum.Enum

Enum representing the different models available.

ag_automm = 'ag_automm'

cat = 'cat'

fasttext = 'fasttext'

gbm = 'gbm'

knn = 'knn'

lr = 'lr'

nn_fastainn = 'nn_fastainn'

nn_mxnet = 'nn_mxnet'

nn_torch = 'nn_torch'

polynomial_linear = 'polynomial_linear'

rf = 'rf'

tabpfn = 'tabpfn'

xgb_linear = 'xgb_linear'

xgb_tree = 'xgb_tree'

xt = 'xt'

actableai.models.autogluon.params.cat module

class actableai.models.autogluon.params.cat.CATParams

Bases: actableai.models.autogluon.params.base.BaseParams

Parameter class for CAT Model.

explain_samples_supported: bool = True

supported_problem_types: Literal['regression', 'binary', 'multiclass', 'quantile'] = ['regression', 'binary', 'multiclass']

actableai.models.autogluon.params.fastainn module

class actableai.models.autogluon.params.fastainn.FastAINNParams

Bases: actableai.models.autogluon.params.base.BaseParams

Parameter class for fastai Model.

explain_samples_supported: bool = False

supported_problem_types: Literal['regression', 'binary', 'multiclass', 'quantile'] = ['regression', 'quantile', 'binary', 'multiclass']

actableai.models.autogluon.params.fasttext module

class actableai.models.autogluon.params.fasttext.FASTTEXTParams

Bases: actableai.models.autogluon.params.base.BaseParams

Parameter class for FASTTEXT Model.

explain_samples_supported: bool = False

supported_problem_types: Literal['regression', 'binary', 'multiclass', 'quantile'] = ['binary', 'multiclass']

actableai.models.autogluon.params.gbm module

class actableai.models.autogluon.params.gbm.GBMParams

Bases: actableai.models.autogluon.params.base.BaseParams

Parameter class for GBM Model.

explain_samples_supported: bool = True

supported_problem_types: Literal['regression', 'binary', 'multiclass', 'quantile'] = ['regression', 'binary', 'multiclass']

actableai.models.autogluon.params.knn module

class actableai.models.autogluon.params.knn.KNNParams

Bases: actableai.models.autogluon.params.base.BaseParams

Parameter class for KNN Model.

explain_samples_supported: bool = False

supported_problem_types: Literal['regression', 'binary', 'multiclass', 'quantile'] = ['regression', 'binary', 'multiclass']

actableai.models.autogluon.params.lr module

class actableai.models.autogluon.params.lr.LRParams

Bases: actableai.models.autogluon.params.base.BaseParams

Parameter class for LR Model.

explain_samples_supported: bool = False

supported_problem_types: Literal['regression', 'binary', 'multiclass', 'quantile'] = ['regression', 'binary', 'multiclass']

actableai.models.autogluon.params.nn_base module

actableai.models.autogluon.params.nn_base.get_parameters() → List[actableai.parameters.base.BaseParameter]

Define parameters that are common for all variants of NN models (PyTorch and MXNet)

Parameters

problem_type – Defines the type of the problem (e.g. regression, binary classification, etc.). See nn_torch.supported_problem_types and nn_mxnet.supported_problem_types for list of accepted strings

Returns

list containing the defined parameters

Return type

parameters

actableai.models.autogluon.params.nn_mxnet module

class actableai.models.autogluon.params.nn_mxnet.NNMXNetParams

Bases: actableai.models.autogluon.params.base.BaseParams

Parameter class for NN Model.

explain_samples_supported: bool = False

supported_problem_types: Literal['regression', 'binary', 'multiclass', 'quantile'] = ['regression', 'binary', 'multiclass']

actableai.models.autogluon.params.nn_torch module

class actableai.models.autogluon.params.nn_torch.NNTorchParams

Bases: actableai.models.autogluon.params.base.BaseParams

Parameter class for NN Model.

explain_samples_supported: bool = False

supported_problem_types: Literal['regression', 'binary', 'multiclass', 'quantile'] = ['quantile']

actableai.models.autogluon.params.polynomial_linear module

class actableai.models.autogluon.params.polynomial_linear.PolynomialLinearPredParams

Bases: actableai.models.autogluon.params.base.BaseParams

Parameter class for Polynomial Linear Predictor Model.

causal_inference_supported = True

explain_samples_supported: bool = True

classmethod get_autogluon_name() → Union[str, Type[AbstractModel]]

supported_problem_types: Literal['regression', 'binary', 'multiclass', 'quantile'] = ['regression']

actableai.models.autogluon.params.rf module

class actableai.models.autogluon.params.rf.RFParams

Bases: actableai.models.autogluon.params.base.BaseParams

Parameter class for RF Model.

explain_samples_supported: bool = True

supported_problem_types: Literal['regression', 'binary', 'multiclass', 'quantile'] = ['regression', 'quantile', 'binary', 'multiclass']

actableai.models.autogluon.params.tabpfn module

class actableai.models.autogluon.params.tabpfn.TabPFNParams

Bases: actableai.models.autogluon.params.base.BaseParams

Parameter class for TabPFN Model.

explain_samples_supported: bool = False

classmethod get_autogluon_name() → Union[str, Type[AbstractModel]]

supported_problem_types: Literal['regression', 'binary', 'multiclass', 'quantile'] = ['binary', 'multiclass']

actableai.models.autogluon.params.xgboost_base module

actableai.models.autogluon.params.xgboost_base.get_parameters(problem_type: str, num_class: int) → List[actableai.parameters.base.BaseParameter]

Define parameters that are common for all variants of XGBoost boosters (tree and linear)

Parameters

• problem_type – Defines the type of the problem (e.g. regression, binary classification, etc.). See supported_problem_types variable for list of accepted strings
• num_class – The number of classes, used for multi-class classification.

Returns

list containing the defined parameters

Return type

parameters

actableai.models.autogluon.params.xgboost_linear module

class actableai.models.autogluon.params.xgboost_linear.XGBoostLinearParams

Bases: actableai.models.autogluon.params.base.BaseParams

Parameter class for XGBoost Model when using a linear booster.

explain_samples_supported: bool = True

supported_problem_types: Literal['regression', 'binary', 'multiclass', 'quantile'] = []

actableai.models.autogluon.params.xgboost_tree module

class actableai.models.autogluon.params.xgboost_tree.XGBoostTreeParams

Bases: actableai.models.autogluon.params.base.BaseParams

Parameter class for XGBoost Model when using a tree booster.

explain_samples_supported: bool = True

supported_problem_types: Literal['regression', 'binary', 'multiclass', 'quantile'] = ['regression', 'binary', 'multiclass']

actableai.models.autogluon.params.xt module

class actableai.models.autogluon.params.xt.XTParams

Bases: actableai.models.autogluon.params.base.BaseParams

Parameter class for XT Model.

explain_samples_supported: bool = True

supported_problem_types: Literal['regression', 'binary', 'multiclass', 'quantile'] = ['regression', 'binary', 'multiclass']

Module contents