Sklearn XGBoost Regressor

XGBoost Regressor for regression tasks. It implements gradient boosted decision trees designed for speed and performance.

Type

ml-estimator

Class

fire.nodes.sklearn.NodeXGBoostRegressor

Fields

Name

Title

Description

targetCol

Target Column

The label column for model fitting

featureCols

Feature Columns

Feature columns of type - all numeric, boolean and vector

splitRatio

Split Ratio

Split Ratio

n_estimators

Number of Estimators

Number of boosting rounds (trees).

max_depth

Max Depth

Maximum depth of a tree. Increasing makes model more complex.

learning_rate

Learning Rate

Boosting learning rate (xgb’s eta).

subsample

Subsample

Subsample ratio of the training instance.

colsample_bytree

Colsample Bytree

Subsample ratio of columns when constructing each tree.

colsample_bylevel

Colsample Bylevel

Subsample ratio of columns for each split, in each level.

colsample_bynode

Colsample Bynode

Subsample ratio of columns for each split, in each node.

reg_alpha

Reg Alpha (L1)

L1 regularization term on weights.

reg_lambda

Reg Lambda (L2)

L2 regularization term on weights.

gamma

Gamma

Minimum loss reduction required to make a further partition on a leaf node.

min_child_weight

Min Child Weight

Minimum sum of instance weight needed in a child.

max_delta_step

Max Delta Step

Maximum delta step allowed for each tree’s weight estimation.

random_state

Random State

Random number seed.

n_jobs

Number of Jobs

Number of parallel threads used to run xgboost.

verbosity

Verbosity

Verbosity of printing messages (0 = silent, 1 = warning, 2 = info, 3 = debug).

booster

Booster

Specify which booster to use.

tree_method

Tree Method

Tree construction algorithm used in XGBoost.

objective

Objective

Learning task and objective function.

eval_metric

Evaluation Metric

Evaluation metric for validation data.

early_stopping_rounds

Early Stopping Rounds

Activates early stopping. Validation metric needs to improve at least once in every given rounds.