mjoint.object: Fitted 'mjoint' object
In joineRML: Joint Modelling of Multivariate Longitudinal Data and Time-to-Event Outcomes
mjoint.object R Documentation
Fitted mjoint object
Description
An object returned by the mjoint function, inheriting
from class mjoint and representing a fitted joint model for
multivariate longitudinal and time-to-event data. Objects of this class
have methods for the generic functions coef, logLik,
plot, print, ranef, fixef, summary,
AIC, getVarCov, vcov, confint, sigma,
fitted, residuals, and formula.
Usage
mjoint.object
Format
An object of class NULL of length 0.
Value
A list with the following components.
coefficientsa list with the estimated coefficients. The
components of this list are:
betathe vector of fixed effects for the linear mixed
effects sub-model.
Dthe variance-covariance matrix of the random effects.
sigma2the measurement error standard deviations for the
linear mixed effects sub-model.
hazthe estimated baseline hazard values for each unique
failure time. Note that this is the centered hazard, equivalent to
that returned by coxph.detail.
gammathe vector of baseline covariates for the survival
model and the latent association coefficient parameter estimates.
historya matrix with parameter estimates at each iteration
of the MCEM algorithm.
nMC.hxa vector with the number of Monte Carlo samples for
each MCEM algorithm iteration.
formLongFixeda list of formulae for the fixed effects
component of each longitudinal outcome.
formLongRandoma list of formulae for the fixed effects
component of each longitudinal outcome. The length of the list will be
equal to formLongFixed.
formSurva formula specifying the proportional hazards
regression model (not including the latent association structure).
dataa list of data.frames for each longitudinal outcome.
survDataa data.frame of the time-to-event dataset.
timeVara character string vector of length K denoting the
column name(s) for time in data.
ida character string denoting the column name for subject
IDs in data and survData.
dimsa list giving the dimensions of model parameters with
components:
pa vector of the number of fixed effects for each
longitudinal outcome.
ra vector of the number of random effects for each
longitudinal outcome.
Kan integer of the number of different longitudinal
outcome types.
qan integer of the number of baseline covariates in the
time-to-event sub-model.
nan integer of the total number of subjects in the study.
nka vector of the number of measurements for each
longitudinal outcome.
sfitan object of class coxph for the separate
time-to-event model fit. See coxph for details.
lfita list of objects each of class lme from fitting
separate linear mixed effects models; one per each longitudinal outcome
type. See lme for details.
log.lik0the combined log-likelihood from separate sub-model
fits.
log.likthe log-likelihood from the joint model fit.
ll.hxa vector of the log-likelihood values for each MCEM
algorithm interaction.
controla list of control parameters used in the estimation
of the joint model. See mjoint for details.
finalnMCthe final number of Monte Carlo samples required
prior to convergence.
callthe matched call.
initsthe initial values passed as an argument in the
mjoint function.
inits.longthe computed initial values from fitting a
multivariate longitudinal model.
inits.survthe computed initial values from fitting a Cox
proportional hazards model with time-dependent covariates calculated from
the fitted multivariate LME model.
convlogical: did the MCEM algorithm converge within the
specified maximum number of iterations?
comp.timea vector of length 2 with each element an object of
class difftime that reports the total time taken for model
fitting (including all stages) and the time spent in the EM
algorithm.
Post model fit statistics
If pfs=TRUE, indicating that post-fit statistics are to be returned,
then the output also includes the following objects.
vcovthe variance-covariance matrix of model parameters, as
approximated by the empirical information matrix, is reported. See
mjoint for details.
SE.approxthe square-root of the diagonal of vcov is
returned, which are estimates of the standard errors for the parameters.
Eba matrix with the estimated random effects values for
each subject.
Vban array with the estimated variance-covariance matrices
for the random effects values for each subject.
dmatsa list of length 3 containing the design matrices,
data frames, and vectors used in the MCEM algorithm. These are required
for prediction and to calculate the residuals and . The 3 items in the
list are l (longitudinal data), t (time-to-event data), and
z (design matrices expanded over unique failure times). These are
not intended to be extracted by the user.
Author(s)
Graeme L. Hickey (graemeleehickey@gmail.com)
See Also
mjoint.
joineRML documentation built on Jan. 22, 2023, 1:18 a.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
| mjoint.object | R Documentation |
Fitted mjoint object
Description
An object returned by the mjoint function, inheriting
from class mjoint and representing a fitted joint model for
multivariate longitudinal and time-to-event data. Objects of this class
have methods for the generic functions coef, logLik,
plot, print, ranef, fixef, summary,
AIC, getVarCov, vcov, confint, sigma,
fitted, residuals, and formula.
Usage
mjoint.object
Format
An object of class NULL of length 0.
Value
A list with the following components.
coefficientsa list with the estimated coefficients. The components of this list are:
betathe vector of fixed effects for the linear mixed effects sub-model.
Dthe variance-covariance matrix of the random effects.
sigma2the measurement error standard deviations for the linear mixed effects sub-model.
hazthe estimated baseline hazard values for each unique failure time. Note that this is the centered hazard, equivalent to that returned by
coxph.detail.gammathe vector of baseline covariates for the survival model and the latent association coefficient parameter estimates.
historya matrix with parameter estimates at each iteration of the MCEM algorithm.
nMC.hxa vector with the number of Monte Carlo samples for each MCEM algorithm iteration.
formLongFixeda list of formulae for the fixed effects component of each longitudinal outcome.
formLongRandoma list of formulae for the fixed effects component of each longitudinal outcome. The length of the list will be equal to
formLongFixed.formSurva formula specifying the proportional hazards regression model (not including the latent association structure).
dataa list of data.frames for each longitudinal outcome.
survDataa data.frame of the time-to-event dataset.
timeVara character string vector of length K denoting the column name(s) for time in
data.ida character string denoting the column name for subject IDs in
dataandsurvData.dimsa list giving the dimensions of model parameters with components:
pa vector of the number of fixed effects for each longitudinal outcome.
ra vector of the number of random effects for each longitudinal outcome.
Kan integer of the number of different longitudinal outcome types.
qan integer of the number of baseline covariates in the time-to-event sub-model.
nan integer of the total number of subjects in the study.
nka vector of the number of measurements for each longitudinal outcome.
sfitan object of class
coxphfor the separate time-to-event model fit. Seecoxphfor details.lfita list of objects each of class
lmefrom fitting separate linear mixed effects models; one per each longitudinal outcome type. Seelmefor details.log.lik0the combined log-likelihood from separate sub-model fits.
log.likthe log-likelihood from the joint model fit.
ll.hxa vector of the log-likelihood values for each MCEM algorithm interaction.
controla list of control parameters used in the estimation of the joint model. See
mjointfor details.finalnMCthe final number of Monte Carlo samples required prior to convergence.
callthe matched call.
initsthe initial values passed as an argument in the
mjointfunction.inits.longthe computed initial values from fitting a multivariate longitudinal model.
inits.survthe computed initial values from fitting a Cox proportional hazards model with time-dependent covariates calculated from the fitted multivariate LME model.
convlogical: did the MCEM algorithm converge within the specified maximum number of iterations?
comp.timea vector of length 2 with each element an object of class
difftimethat reports the total time taken for model fitting (including all stages) and the time spent in the EM algorithm.
Post model fit statistics
If pfs=TRUE, indicating that post-fit statistics are to be returned,
then the output also includes the following objects.
vcovthe variance-covariance matrix of model parameters, as approximated by the empirical information matrix, is reported. See
mjointfor details.SE.approxthe square-root of the diagonal of
vcovis returned, which are estimates of the standard errors for the parameters.Eba matrix with the estimated random effects values for each subject.
Vban array with the estimated variance-covariance matrices for the random effects values for each subject.
dmatsa list of length 3 containing the design matrices, data frames, and vectors used in the MCEM algorithm. These are required for prediction and to calculate the residuals and . The 3 items in the list are
l(longitudinal data),t(time-to-event data), andz(design matrices expanded over unique failure times). These are not intended to be extracted by the user.
Author(s)
Graeme L. Hickey (graemeleehickey@gmail.com)
See Also
mjoint.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.