BoostLGBM

class lightautoml.ml_algo.boost_lgbm.BoostLGBM(default_params=None, freeze_defaults=True, timer=None, optimization_search_space={})[source]

Bases: TabularMLAlgo, ImportanceEstimator

Gradient boosting on decision trees from LightGBM library.

default_params: All available parameters listed in lightgbm documentation:

https://lightgbm.readthedocs.io/en/latest/Parameters.html

freeze_defaults:

True : params may be rewritten depending on dataset.

False: params may be changed only manually or with tuning.

timer: Timer instance or None.

init_params_on_input(train_valid_iterator)[source]

Get model parameters depending on dataset parameters.

Parameters:: train_valid_iterator (TrainValidIterator) – Classic cv-iterator.
Return type:: dict
Returns:: Parameters of model.

fit_predict_single_fold(train, valid)[source]

Implements training and prediction on single fold.

Parameters:

train (Union[NumpyDataset, CSRSparseDataset, PandasDataset]) – Train Dataset.
valid (Union[NumpyDataset, CSRSparseDataset, PandasDataset]) – Validation Dataset.

Return type:

Tuple[Booster, ndarray]

Returns:

Tuple (model, predicted_values)

predict_single_fold(model, dataset)[source]

Predict target values for dataset.

Parameters:

model (Booster) – Lightgbm object.
dataset (Union[NumpyDataset, CSRSparseDataset, PandasDataset]) – Test Dataset.

Return type:

ndarray

Returns:

Predicted target values.

get_features_score()[source]

Computes feature importance as mean values of feature importance provided by lightgbm per all models.

Return type:: Series
Returns:: Series with feature importances.

fit(train_valid)[source]

Just to be compatible with ImportanceEstimator.

Parameters:: train_valid (TrainValidIterator) – Classic cv-iterator.