CopulaCenR (version 1.2.3)

IRsurv: An information ratio-based goodness-of-fit test for copula models on censored data

Description

Fits an Information ratio (IR)-based goodness-of-fit test for copula models under various censoring types.

Usage

IRsurv(
  data,
  censoring = "rc",
  copula = "clayton",
  fams = c(3, 3, 3),
  R = 200,
  parallel = "no",
  ncpus = 1
)

Value

the p value of the IR test

Arguments

data

the input data; see examples for details.

censoring

types of censoring, including "rc", "ic", "rec_bivariate", "rec_multivariate".

copula

specify the copula family to be tested; default is "clayton"; others include "copula2", "gumbel", "frank", "gaussian", and "d-vine". "d-vine" is only for censoring = "rec_multivariate".

fams

specify the unconditional copulas by following the style of the VineCopula package when copula = "d-vine". Only d = 4 is supported at this stage. The conditional copulas are set as "frank" by default.

R

number of Bootstraps; default is 200.

parallel

indicator of parallel computing; can be "no" or "multicore"; default is "no".

ncpus

number of cpus to be assigned for parallel computing; default is 1.

Examples

if (FALSE) {
# Goodness of fit under right censoring
data("data_sim_RC")
test_rc <- IRsurv(data = data_sim_RC, censoring = "rc", copula = "clayton", R = 200)
test_rc
# Goodness of fit under interval censoring
data("data_sim_ic")
test_ic <- IRsurv(data = data_sim_ic, censoring = "ic", copula = "clayton", R = 200)
test_ic
# Goodness of fit under bivariate recurrent events
data("data_sim_rec")
test_rec_bi <- IRsurv(data = data_sim_rec, censoring = "rec_bivariate",
                      copula = "clayton", R = 200)
test_rec_bi
# Goodness of fit of D-vine copula under multivariate recurrent events
data("data_sim_multi_rec")
test_rec_mv <- IRsurv(data = data_sim_multi_rec, censoring = "rec_multivariate",
                      fams = c(3,3,3), R = 200)
test_rec_mv
}