BARTHyperparameters

Properties

num_trees

integer | object≥ 1

Number of trees in the mean forest.

alpha

number | object> 0< 1

Base of the tree split prior alpha * (1 + depth)^-beta.

beta

number | object≥ 0

Depth penalty exponent of the tree split prior alpha * (1 + depth)^-beta.

min_samples_leaf

integer | object≥ 1

Minimum number of training cases in a leaf. Cannot exceed the number of cases in the training data.

max_depth

integer | object | null≥ 1

Maximum depth of any tree. NULL imposes no limit.

num_features_subsample

integer | object | null≥ 1

Number of features subsampled when growing each tree. NULL uses every feature. Cannot exceed the number of features in the training data.

variance_forest_num_trees

integer | object≥ 0

Number of trees in the conditional variance forest. 0 fits a homoskedastic model.

num_gfr

integer≥ 0

Number of grow-from-root warm-start iterations.

num_burnin

integer≥ 0

Number of burn-in MCMC iterations.

num_mcmc

integer≥ 1

Number of retained MCMC iterations per chain.

num_chains

integer≥ 1

Number of independent MCMC chains. Cannot exceed num_gfr unless num_gfr is 0.

keep_every

integer≥ 1

Thinning interval: retain one MCMC sample in every keep_every.

cutpoint_grid_size

integer≥ 1

Maximum number of candidate cutpoints considered by the grow-from-root algorithm.

standardize

boolean

Center and scale the outcome before sampling.

link

string

Link function of the binary outcome model (classification only). "cloglog" cannot be combined with case weights.

one of"probit""cloglog"

seed

integer | null

Random seed for the sampler. NULL leaves the sampler seeded by the system.

ifw

boolean | object

Inverse Frequency Weighting in classification.

Relationships