softtbart1: Type I Tobit Soft Bayesian Additive Regression Trees with...

In EoghanONeill/TobitBART: Tobit Bayesian Additive Regression Trees

Type I Tobit Soft Bayesian Additive Regression Trees with sparsity inducing hyperprior implemented using MCMC

Description

Type I Tobit Soft Bayesian Additive Regression Trees with sparsity inducing hyperprior implemented using MCMC

Usage

softtbart1(
  x.train,
  x.test,
  y,
  n.iter = 1000,
  n.burnin = 100,
  below_cens = 0,
  above_cens = Inf,
  n.trees = 50L,
  SB_group = NULL,
  SB_alpha = 1,
  SB_beta = 2,
  SB_gamma = 0.95,
  SB_k = 2,
  SB_sigma_hat = NULL,
  SB_shape = 1,
  SB_width = 0.1,
  SB_alpha_scale = NULL,
  SB_alpha_shape_1 = 0.5,
  SB_alpha_shape_2 = 1,
  SB_tau_rate = 10,
  SB_num_tree_prob = NULL,
  SB_temperature = 1,
  SB_weights = NULL,
  SB_normalize_Y = TRUE,
  print.opt = 100,
  fast = TRUE,
  censsigprior = FALSE,
  lambda0 = NA,
  sigest = NA,
  nolinregforsigest = FALSE
)

Arguments

x.train	The training covariate data for all training observations. Number of rows equal to the number of observations. Number of columns equal to the number of covariates.
x.test	The test covariate data for all test observations. Number of rows equal to the number of observations. Number of columns equal to the number of covariates.
y	The training data vector of outcomes. A continuous, censored outcome variable.
n.iter	Number of iterations excluding burnin.
n.burnin	Number of burnin iterations.
below_cens	Number at or below which observations are censored.
above_cens	Number at or above which observations are censored.
n.trees	A positive integer giving the number of trees used in the sum-of-trees formulation.
print.opt	Print every print.opt number of Gibbs samples.
fast	If equal to TRUE, then implements faster truncated normal draws and approximates normal pdf.

Value

The following objects are returned:

Z.matcens	Matrix of draws of latent (censored) outcomes for censored observations. Number of rows equals number of censored training observations. Number of columns equals n.iter . Rows are ordered in order of censored observations in the training data.
Z.matcensbelow	Matrix of draws of latent (censored) outcomes for observations censored from below. Number of rows equals number of training observations censored from below. Number of columns equals n.iter . Rows are ordered in order of censored observations in the training data.
Z.matcensabove	Matrix of draws of latent (censored) outcomes for observations censored from above. Number of rows equals number of training observations censored from above. Number of columns equals n.iter . Rows are ordered in order of censored observations in the training data.
mu	Matrix of draws of the sum of terminal nodes, i.e. f(x_i), for all training observations. Number of rows equals number of training observations. Number of columns equals n.iter .
mucens	Matrix of draws of the sum of terminal nodes, i.e. f(x_i), for all censored training observations. Number of rows equals number of censored training observations. Number of columns equals n.iter .
muuncens	Matrix of draws of the sum of terminal nodes, i.e. f(x_i), for all uncensored training observations. Number of rows equals number of uncensored training observations. Number of columns equals n.iter .
mucensbelow	Matrix of draws of the sum of terminal nodes, i.e. f(x_i), for all training observations censored from below. Number of rows equals number of training observations censored from below. Number of columns equals n.iter .
mucensabove	Matrix of draws of the sum of terminal nodes, i.e. f(x_i), for all training observations censored from above Number of rows equals number of training observations censored from above Number of columns equals n.iter .
ystar	Matrix of training sample draws of the outcome assuming uncensored (can take values below below_cens and above above_cens. Number of rows equals number of training observations. Number of columns equals n.iter .
ystarcens	Matrix of censored training sample draws of the outcome assuming uncensored (can take values below below_cens and above above_cens. Number of rows equals number of censored training observations. Number of columns equals n.iter .
ystaruncens	Matrix of uncensored training sample draws of the outcome assuming uncensored (can take values below below_cens and above above_cens. Number of rows equals number of uncensored training observations. Number of columns equals n.iter .
ystarcensbelow	Matrix of censored from below training sample draws of the outcome assuming uncensored (can take values below below_cens and above above_cens. Number of rows equals number of training observations censored from below. Number of columns equals n.iter .
ystarcensabove	Matrix of censored from above training sample draws of the outcome assuming uncensored (can take values below below_cens and above above_cens. Number of rows equals number of training observations censored from above. Number of columns equals n.iter .
test.mu	Matrix of draws of the sum of terminal nodes, i.e. f(x_i), for all test observations. Number of rows equals number of test observations. Number of columns equals n.iter .
test.y_nocensoring	Matrix of test sample draws of the outcome assuming uncensored. Can take values below below_cens and above above_cens. Number of rows equals number of test observations. Number of columns equals n.iter .
test.y_withcensoring	Matrix of test sample draws of the outcome assuming censored. Cannot take values below below_cens and above above_cens. Number of rows equals number of test observations. Number of columns equals n.iter .
test.probcensbelow	Matrix of draws of probabilities of test sample observations being censored from below. Number of rows equals number of test observations. Number of columns equals n.iter .
test.probcensabove	Matrix of draws of probabilities of test sample observations being censored from above. Number of rows equals number of test observations. Number of columns equals n.iter .
sigma	Vector of draws of the standard deviation of the error term. Number of elements equals n.iter .

Examples

#example taken from https://stats.idre.ucla.edu/r/dae/tobit-models/

dat <- read.csv("https://stats.idre.ucla.edu/stat/data/tobit.csv")

train_inds <- sample(1:200,190)
test_inds <- (1:200)[-train_inds]

ytrain <- dat$apt[train_inds]
ytest <- dat$apt[test_inds]

xtrain <- cbind(dat$read, dat$math)[train_inds,]
xtest <- cbind(dat$read, dat$math)[test_inds,]

tobart_res <- tbart1(xtrain,xtest,ytrain,
                    below_cens = -Inf,
                    above_cens = 800,
                    n.iter = 400,
                    n.burnin = 100)

EoghanONeill/TobitBART documentation built on Feb. 6, 2025, 6:52 a.m.