LinearLBFGS

class lightautoml.ml_algo.linear_sklearn.LinearLBFGS(default_params=None, freeze_defaults=True, timer=None, optimization_search_space={})[source]

Bases: lightautoml.ml_algo.base.TabularMLAlgo

LBFGS L2 regression based on torch.

default_params:

  • cs: List of regularization coefficients.

  • max_iter: Maximum iterations of L-BFGS.

  • tol: The tolerance for the stopping criteria.

  • early_stopping: Maximum rounds without improving.

freeze_defaults:

  • True : params may be rewrited depending on dataset.

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

timer: Timer instance or None.

fit_predict_single_fold(train, valid)[source]

Train on train dataset and predict on holdout dataset.

Parameters
Return type

Tuple[TorchBasedLinearEstimator, ndarray]

Returns

Target predictions for valid dataset.

predict_single_fold(model, dataset)[source]

Implements prediction on single fold.

Parameters
Return type

ndarray

Returns

Predictions for input dataset.