LightRuleFitHyperparameters
VariantLightRuleFit (LightGBM rules + GLMNET). See `setup_LightRuleFit`.
Properties
nroundstunableinteger≥ 1Number of boosting rounds (LightGBM step).
num_leavestunableinteger≥ 1Maximum number of leaves in one tree (LightGBM step).
max_depthtunableintegerMaximum tree depth (LightGBM step). -1 = no limit.
learning_ratetunablenumber≤ 1> 0Learning rate (LightGBM step).
subsampletunablenumber≤ 1> 0Fraction of cases sampled per tree (LightGBM step).
subsample_freqtunableinteger≥ 1Bagging frequency (LightGBM step).
lambda_l1tunablenumber≥ 0L1 regularization (LightGBM step).
lambda_l2tunablenumber≥ 0L2 regularization (LightGBM step).
objectivestring | nullLightGBM objective. NULL = set from outcome type.
ifw_lightgbmtunablebooleanInverse Frequency Weighting in the LightGBM step.
alphatunablenumber≥ 0≤ 1Elastic net mixing parameter (GLMNET step).
lambdaarray | null≥ 1 itemsRegularization strength (GLMNET step). NULL = determined by cv.glmnet.
ifw_glmnettunablebooleanInverse Frequency Weighting in the GLMNET step.
ifwbooleanInverse Frequency Weighting in both steps. Cannot be combined with ifw_lightgbm or ifw_glmnet.
Relationships
Used by