loglik: Wrapper to the Fortran subroutines computing the...

In CecileProust-Lima/lcmm: Extended Mixed Models Using Latent Classes and Latent Processes

Wrapper to the Fortran subroutines computing the log-likelihood

Description

Log-likelihood of hlme, lcmm, multlcmm, Jointlcmm and mpjlcmm models. The argument's specification is not straightforward, so these functions are usually not directly used.

Usage

loglikhlme(
  b,
  Y0,
  X0,
  prior0,
  pprior0,
  idprob0,
  idea0,
  idg0,
  idcor0,
  ns0,
  ng0,
  nv0,
  nobs0,
  nea0,
  nmes0,
  idiag0,
  nwg0,
  ncor0,
  npm0,
  fix0,
  nfix0,
  bfix0
)

logliklcmm(
  b,
  Y0,
  X0,
  prior0,
  pprior0,
  idprob0,
  idea0,
  idg0,
  idcor0,
  ns0,
  ng0,
  nv0,
  nobs0,
  nea0,
  nmes0,
  idiag0,
  nwg0,
  ncor0,
  npm0,
  epsY0,
  idlink0,
  nbzitr0,
  zitr0,
  minY0,
  maxY0,
  ide0,
  fix0,
  nfix0,
  bfix0
)

loglikmultlcmm(
  b,
  Y0,
  X0,
  prior0,
  pprior0,
  idprob0,
  idea0,
  idg0,
  idcor0,
  idcontr0,
  ny0,
  ns0,
  ng0,
  nv0,
  nobs0,
  nea0,
  nmes0,
  idiag0,
  nwg0,
  ncor0,
  nalea0,
  npm0,
  epsY0,
  idlink0,
  nbzitr0,
  zitr0,
  uniqueY0,
  indiceY0,
  nvalSPLORD0,
  fix0,
  nfix0,
  bfix0,
  methInteg0,
  nMC0,
  dimMC0,
  seqMC0,
  chol0
)

loglikJointlcmm(
  b,
  Y0,
  X0,
  prior0,
  pprior0,
  tentr0,
  tevt0,
  devt0,
  ind_survint0,
  idprob0,
  idea0,
  idg0,
  idcor0,
  idcom0,
  idspecif0,
  idtdv0,
  idlink0,
  epsY0,
  nbzitr0,
  zitr0,
  uniqueY0,
  nvalSPL0,
  indiceY0,
  typrisq0,
  risqcom0,
  nz0,
  zi0,
  ns0,
  ng0,
  nv0,
  nobs0,
  nmes0,
  nbevt0,
  nea0,
  nwg0,
  ncor0,
  idiag0,
  idtrunc0,
  logspecif0,
  npm0,
  fix0,
  nfix0,
  bfix0
)

loglikmpjlcmm(
  b,
  K0,
  ny0,
  nbevt0,
  ng0,
  ns0,
  Y0,
  nobs0,
  X0,
  nv0,
  Xns0,
  nv20,
  prior0,
  Tentr0,
  Tevt0,
  Devt0,
  ind_survint0,
  idnv0,
  idnv20,
  idspecif0,
  idlink0,
  epsY0,
  nbzitr0,
  zitr0,
  uniqueY0,
  nvalSPL0,
  indiceY0,
  typrisq0,
  risqcom0,
  nz0,
  zi0,
  nmes0,
  nea0,
  nw0,
  ncor0,
  nalea0,
  idiag0,
  idtrunc0,
  logspecif0,
  npm0,
  fix0,
  contrainte0,
  nfix0,
  bfix0
)

Arguments

b	the vector of estimated parameters (length npm0)
Y0	the observed values of the outcome(s) (length nobs0)
X0	the observed values of all covariates included in the model (dim nob0 * nv0)
prior0	the prior latent class (length ns0)
pprior0	the prior probabilty of each latent class (dim ns0 * ng0)
idprob0	indicator of presence in the class membership submodel (length nv0)
idea0	indicator of presence in the random part of the longitudinal submodel (length nv0)
idg0	indicator of presence in the fixed part of the longitudinal submodel (length nv0)
idcor0	indicator of presence in the correlation part of the longitudinal submodel (length nv0)
ns0	number of subjects
ng0	number of latent classes
nv0	number of covariates
nobs0	number of observations
nea0	number of random effects
nmes0	number of mesures for each subject (length ns0 or dom ns0*ny0)
idiag0	indicator of diagonal variance matrix of the random effects
nwg0	number of parameters for proportional random effects over latent classes
ncor0	number of parameters for the correlation
npm0	total number of parameters
fix0	indicator of non estimated parameter (length npm0+nfix0)
nfix0	number of non estimated parameters
bfix0	vector of non estimated parameters
epsY0	epsY values for Beta transformations
idlink0	type of transformation
nbzitr0	number of nodes for the transformations
zitr0	nodes for the transformations
minY0	minimum value for the longitudinal outcome
maxY0	maximum value for the longitudinal outcome
ide0	indicator of observed values for ordinal outcomes
idcontr0	indicator of presence as contrast in the fixed part of the longitudinal submodel (length nv0)
ny0	number of longitudinal outcomes
nalea0	number of parameters f the outcome specific random effect
uniqueY0	unique values of the longitudinal outcomes
indiceY0	correspondance between Y0 and uniqueY0
nvalSPLORD0	number of unique values for outcomes modeled with splines transformations or as ordinal outcome
methInteg0	type of integration
nMC0	number of nodes for Monte Carlo integration
dimMC0	dimension of the integration
seqMC0	sequence of integration nodes
chol0	indicator of Cholesky parameterization
tentr0	entry time for the survival submodel
tevt0	event time for the survival submodel
devt0	indicator of event for the survival submodel
ind_survint0	indicator of risk change
idcom0	indicator of presence in the survival submodel with common effect
idspecif0	indicator of presence in the survival submodel with cause-specific or class specific effect
idtdv0	indicator of 'TimeDepVar' covariate
nvalSPL0	number of unique values for outcomes modeled with splines transformations
typrisq0	type of baseline risk
risqcom0	specification of baseline risk across latent classes
nz0	number of nodes for the baseline
zi0	nodes for the baseline
nbevt0	number of events
idtrunc0	indicator of left truncation
logspecif0	indicator of logarithm parameterization
K0	number of latent processes
Xns0	the observed values of the covariates included in the survival submodel (dim ns0*nv20)
nv20	number of covariates in Xns0
Tentr0	entry time for the survival submodel (length ns0)
Tevt0	event time for the survival submodel (length ns0)
Devt0	indicator of event for the survival submodel (length ns0)
idnv0	indicator of presence in each subpart of the longitudinal models (length 4*sum(nv0))
idnv20	indicator of presence in each subpart of the survival models (length 3*nv20)
nw0	number of parameters for proportional random effects over latent classes
contrainte0	type of identifiability constraints

Value

the log-likelihood

Author(s)

Cecile Proust-Lima, Viviane Philipps

CecileProust-Lima/lcmm documentation built on March 3, 2024, 5:23 p.m.