SparkLinearLBFGS

class sparklightautoml.ml_algo.linear_pyspark.SparkLinearLBFGS(default_params=None, freeze_defaults=True, timer=None, optimization_search_space=None, persist_output_dataset=True, computations_settings=None)[source]

Bases: SparkTabularMLAlgo

LBFGS L2 regression based on Spark MLlib.

default_params:

  • tol: The tolerance for the stopping criteria.

  • maxIter: Maximum iterations of L-BFGS.

  • aggregationDepth: Param for suggested depth for treeAggregate.

  • elasticNetParam: Elastic net parameter.

  • regParam: Regularization parameter.

  • 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.

predict_single_fold(dataset, model)[source]

Implements prediction on single fold.

Parameters:
  • model (PipelineModel) – Model uses to predict.

  • dataset (SparkDataset) – SparkDataset used for prediction.

Return type:

DataFrame

Returns:

Predictions for input dataset.