joint.object {gmvjoint}

Fitted joint object

Description

An object returned by the joint function, with class joint a fitted joint model. Objects of this class currently have methods for: logLik, print, ranef, fixef, summary, AIC, and vcov.

Usage

joint.object

Format

An object of class NULL of length 0.

Value

A list with the following components.

coeffs

A list containing parameter estimates:

D

The variance-covariance matrix of the random effects.

beta

Vector of fixed effects for longitudinal processes.

sigma

List of dispersion parameters, families with no dispersion parameter are returned as an unnamed zero value.

gamma

Vector of association parameters.

zeta

Vector of time-invariant survival coefficients.

hazard

A matrix containing unique failure times ft, their hazard contribution haz and the number of events at that failure time nev.

ModelInfo

A list containing information on the model fit:

ResponseInfo

A vector containing response names with (family) reported.

Resps

A vector containing response names only.

family

A list of families fit.

K

An integer specifying the number of longitudinal sub-models.

Pcounts

A list containing informations about the number of parameters/random effects:

P

A vector of length K containing the number of fixed effects for each response (in order).

Pd

A vector of length K containing the number of dispersion parameters for each response (in order) 0 denotes no parameter for that response.

q

An integer denoting the number of random effects.

vD

An integer denoting the number of unique variance-covariance parameters estimated.

long.formulas

A list of long.formulas (i.e. from joint call).

disp.formulas

A list of disp.formulas (i.e. from joint call). If no disp.formulas are supplied to joint, then this is populated by a list of K "~1". The environment is set to parent.frame in this case to avoid memory overheads in returned objects.

surv.formula

Formula object from joint call.

survtime

The name of the event time used in surv.formula.

status

The name of the event indicator used in surv.formula.

control

List of control parameters used, see joint.

convergence.criteria

List of parameters relating to the stopping rule.

inds

A list of length two, named R and Cpp, each of which contains the indices for fixed effects \beta for each response, or the random effects b for the named platform.

n

Number of subjects.

nobs

A vector containing total number of observations for each response.

mi

A K x n matrix containing the number of observations for subject i for the kth response.

nev

Number of events.

id.assign

A list containing the original ids of subjects in the data supplied to joint, and the id assigned to them for use in subsequent functions.

Hessian

The (approximated) Hessian found at MLEs. Only returned if control argument post.process=TRUE

.

vcov

The full variance-covariance matrix between parameters. Only returned if control argument post.process=TRUE.

SE

A named vector of approximated standard error for each estimated parameter. Only returned if control argument post.process=TRUE.

logLik

log-likelihood evaluated at parameter estimates. Only returned if control argument post.process=TRUE.

REs

The random effects, with subject-specific variance matrices attributed. If control argumnet post.process=TRUE then these are found at MLEs (i.e. are posterior estimates), otherwise they are taken from the final EM iteration.

elapsed.time

Named numeric containing breakdown of elapsed time for joint fit.

dmats

A list of data matrices on each of the longitudinal and survival processes for each subject.

Author(s)

James Murray ([e-mail]).

See Also

joint.