xgboost.spark.core Namespace Reference

Variables
list	_pyspark_specific_params
list	_non_booster_params = ["missing", "n_estimators", "feature_types", "feature_weights"]
dict	_pyspark_param_alias_map
dict	_inverse_pyspark_param_alias_map = {v: k for k, v in _pyspark_param_alias_map.items()}
list	_unsupported_xgb_params
dict	_unsupported_fit_params

Detailed Description

XGBoost pyspark integration submodule for core code.

Variable Documentation

_pyspark_param_alias_map

dict xgboost.spark.core._pyspark_param_alias_map

protected

Initial value:

=  {
    "features_col": "featuresCol",
    "label_col": "labelCol",
    "weight_col": "weightCol",
    "raw_prediction_col": "rawPredictionCol",
    "prediction_col": "predictionCol",
    "probability_col": "probabilityCol",
    "validation_indicator_col": "validationIndicatorCol",
}

_pyspark_specific_params

list xgboost.spark.core._pyspark_specific_params

protected

Initial value:

=  [
    "featuresCol",
    "labelCol",
    "weightCol",
    "rawPredictionCol",
    "predictionCol",
    "probabilityCol",
    "validationIndicatorCol",
    "base_margin_col",
    "arbitrary_params_dict",
    "force_repartition",
    "num_workers",
    "feature_names",
    "features_cols",
    "enable_sparse_data_optim",
    "qid_col",
    "repartition_random_shuffle",
    "pred_contrib_col",
    "use_gpu",
]

_unsupported_fit_params

dict xgboost.spark.core._unsupported_fit_params

protected

Initial value:

=  {
    "sample_weight",  # Supported by spark param weightCol
    "eval_set",  # Supported by spark param validation_indicator_col
    "sample_weight_eval_set",  # Supported by spark param weight_col + validation_indicator_col
    "base_margin",  # Supported by spark param base_margin_col
    "base_margin_eval_set",  # Supported by spark param base_margin_col + validation_indicator_col
    "group",  # Use spark param `qid_col` instead
    "qid",  # Use spark param `qid_col` instead
    "eval_group",  # Use spark param `qid_col` instead
    "eval_qid",  # Use spark param `qid_col` instead
}

_unsupported_xgb_params

list xgboost.spark.core._unsupported_xgb_params

protected

Initial value:

=  [
    "gpu_id",  # we have "device" pyspark param instead.
    "enable_categorical",  # Use feature_types param to specify categorical feature instead
    "use_label_encoder",
    "n_jobs",  # Do not allow user to set it, will use `spark.task.cpus` value instead.
    "nthread",  # Ditto
]