BayesERtools (version 0.2.0)

dev_ermod_bin_cov_sel: Perform covariate selection for linear ER model

Description

This functions is used to develop an ER model with covariates for binary and continuous endpoints. projpred package is used for variable selection.

Usage

dev_ermod_bin_cov_sel(
  data,
  var_resp,
  var_exposure,
  var_cov_candidates,
  cv_method = c("LOO", "kfold"),
  k = 5,
  validate_search = FALSE,
  nterms_max = NULL,
  .reduce_obj_size = TRUE,
  verbosity_level = 1,
  chains = 4,
  iter = 2000
)
dev_ermod_lin_cov_sel(
  data,
  var_resp,
  var_exposure,
  var_cov_candidates,
  cv_method = c("LOO", "kfold"),
  k = 5,
  validate_search = FALSE,
  nterms_max = NULL,
  .reduce_obj_size = TRUE,
  verbosity_level = 1,
  chains = 4,
  iter = 2000
)

Value

An object of class ermod_bin_cov_sel or ermod_lin_cov_sel.

Arguments

data

Input data for E-R analysis

var_resp

Response variable name in character

var_exposure

Exposure variable names in character

var_cov_candidates

Candidate covariate names in character vector

cv_method

Cross-validation method. Default is "LOO" (recommended). Use "kfold" if you see warnings on Pareto k estimates.

k

Number of folds for kfold CV. Only used if cv_method is "kfold".

validate_search

Whether to validate the search. Default is FALSE. Recommend to set to TRUE for kfold CV. Do not use for LOO (run time would become too long).

nterms_max

Maximum number of terms to consider in the model. Default is NULL (all terms are considered).

.reduce_obj_size

Whether to reduce object size by removing some elements from projpred outputs that are not necessary for the functionality of this package.

verbosity_level

Verbosity level. 0: No output, 1: Display steps, 2: Display progress in each step, 3: Display MCMC sampling.

chains

Number of chains for Stan.

iter

Number of iterations for Stan.

Examples

# \donttest{
data(d_sim_binom_cov_hgly2)

er_binary_cov_model <- dev_ermod_bin_cov_sel(
  data = d_sim_binom_cov_hgly2,
  var_resp = "AEFLAG",
  var_exposure = "AUCss_1000",
  var_cov_candidates = c(
    "BAGE_10", "BWT_10", "BGLUC",
    "BHBA1C_5", "RACE", "VISC"
  )
)

er_binary_cov_model
# }

# \donttest{
data(d_sim_lin)

ermod_lin_cov_sel <- dev_ermod_lin_cov_sel(
  data = d_sim_lin,
  var_resp = "response",
  var_exposure = "AUCss",
  var_cov_candidates = c("BAGE", "SEX")
)

ermod_lin_cov_sel
# }