depCensoring (version 0.1.7)

cr.lik: Competing risk likelihood function.

Description

This function implements the second step likelihood function of the competing risk model defined in Willems et al. (2024+).

Usage

cr.lik(
  n,
  s,
  Y,
  admin,
  cens.inds,
  M,
  Sigma,
  beta.mat,
  sigma.vct,
  rho.mat,
  theta.vct
)

Value

Evaluation of the log-likelihood function

Arguments

n

The sample size.

s

The number of competing risks.

Y

The observed times.

admin

Boolean value indicating whether or not administrative censoring should be taken into account.

cens.inds

matrix of censoring indicators (each row corresponds to a single observation).

M

Design matrix, consisting of [intercept, exo.cov, Z, cf]. Note that cf represents the multiple ways of 'handling' the endogenous covariate Z, see also the documentation of 'estimate.cmprsk.R'. When there is no confounding, M will be [intercept, exo.cov].

Sigma

The covariance matrix.

beta.mat

Matrix containing all of the covariate effects.

sigma.vct

Vector of standard deviations. Should be equal to sqrt(diag(Sigma)).

rho.mat

The correlation matrix.

theta.vct

Vector containing the parameters of the Yeo-Johnsontrans- formations.

References

Willems et al. (2024+). Flexible control function approach under competing risks (in preparation).