rbartpackages.missBART.Hypers

class rbartpackages.missBART.Hypers

Prior hyperparameters, the keys of R’s hypers_list.

Entries left out fall back to hypers_list’s own defaults, noted below; kappa, alpha and V also accept None to request their data-based default explicitly.

mu0: float

Prior mean of the tree-leaf parameters. Default 0.

kappa: float | None

Prior precision of the regression-tree leaf parameters. Default 4 * qnorm(0.975) ** 2 * n_reg_trees.

alpha: float | None

Wishart degrees of freedom of the residual precision, multivariate y only. Default df.

V: Float64[ndarray, 'p p'] | None

Wishart scale matrix of the residual precision, multivariate y only. Default data-based.

df: float

Degrees of freedom of the error-variance prior. Default 10.

q: float

Prior quantile of the error variance at the data-based estimate. Default 0.75.

tau_b: float

Unused by missBART2. Default 100.