LightGBMHyperparameters

Properties

max_nrounds

integer≥ 1

Maximum number of boosting rounds when tuning nrounds by early stopping.

force_nrounds

integer | null≥ 1

Use this many boosting rounds; disables the search for nrounds.

early_stopping_rounds

integer≥ 1

Number of rounds without improvement to stop training.

num_leavestunable

integer≥ 1

Maximum number of leaves in one tree.

max_depthtunable

integer

Maximum tree depth. -1 = no limit.

learning_ratetunable

number≤ 1> 0

Learning rate.

feature_fractiontunable

number≤ 1> 0

Fraction of features sampled per tree.

subsampletunable

number≤ 1> 0

Fraction of cases sampled per tree (bagging fraction).

subsample_freqtunable

integer≥ 1

Bagging frequency.

lambda_l1tunable

number≥ 0

L1 regularization.

lambda_l2tunable

number≥ 0

L2 regularization.

max_cat_thresholdtunable

integer≥ 1

Maximum number of split points for categorical features.

min_data_per_grouptunable

integer≥ 1

Minimum number of cases per categorical group.

linear_treetunable

boolean

Fit linear models at leaves.

ifwtunable

boolean

Inverse Frequency Weighting in classification.

objective

string | null

LightGBM objective. NULL = set from outcome type.

device_type

string

Compute device.

one of"cpu""gpu""cuda"

tree_learner

string

Tree learner type.

one of"serial""feature""data""voting"

force_col_wise

boolean

Force column-wise histogram building (CPU only).

nrounds

integer | null≥ 1

Resolved number of boosting rounds. NULL = determined by early stopping during tuning.

best_iter

number | null≥ 0

Best iteration found by early stopping.

Relationships