glmtoolbox: glmgee – R documentation

Become an expert in R — Interactive courses, Cheat Sheets, certificates and more!

glmgee

Generalized Estimating Equations

Description

Produces an object of the class glmgee in which are stored the main results of the generalized estimating equation fitted to the data.

Usage

glmgee(
  formula,
  family = gaussian(),
  weights,
  id,
  waves,
  data,
  subset,
  corstr,
  corr,
  start,
  scale.fix = FALSE,
  scale.value,
  toler = 1e-07,
  maxit = 50,
  adjr2 = FALSE,
  ...
)

Arguments

`formula`	a `formula` expression of the form `response ~ predictors`, which is a symbolic description of the linear predictor of the model to be fitted to the data.
`family`	a `family` object, that is, a list of functions and expressions for defining link and variance functions. Families supported are gaussian, binomial, poisson, Gamma, inverse gaussian and quasi. See the glm and family documentation. By default, `family` is set to be `gaussian("identity")`.
`weights`	an (optional) vector of positive "prior weights" to be used in the fitting process. The length of `weights` should be the same as the number of observations.
`id`	a vector which identifies the clusters. The length of `id` should be the same as the number of observations.
`waves`	an (optional) a positive integer-valued variable that is used to identify the order and spacing of observations within groups. This argument is crucial when there are missing values and gaps in the data. By default, `waves` is equal to the integers from 1 to the size of each cluster.
`data`	an (optional) `data frame` in which to look for variables involved in the `formula` expression, as well as the variables specified in the arguments `id` and `weights`. The data are assumed to be sorted by `id` and time.
`subset`	an (optional) vector specifying a subset of observations to be used in the fitting process.
`corstr`	a character string specifying the (working) correlation structure. The available options are: "Independence", "Unstructured", "Stationary-M-dependent(m)", "Non-Stationary-M-dependent(m)", "AR-1", "Exchangeable" and "User-defined", where m represents the lag of the dependence. By default, `corstr` is set to be "Independence".
`corr`	an (optional) square matrix of the same dimension of the maximum cluster size containing the user specified correlation. This is only appropriate if `corstr` is specified to be "User-defined".
`start`	an (optional) vector of starting values for the parameters in the linear predictor.
`scale.fix`	an (optional) logical variable. If TRUE, the scale parameter is fixed at the value of `scale.value`. By default, `scale.fix` is set to be FALSE.
`scale.value`	an (optional) numeric variable giving the value to which the scale parameter should be fixed. This is only appropriate if `scale.fix=TRUE`. By default, `scale.value` is set to be 1.
`toler`	an (optional) positive value which represents the convergence tolerance. The convergence is reached when the maximum of the relative differences between the values of the parameters in the linear predictor in consecutive iterations of the fitting algorithm is lower than `toler`. By default, `toler` is set to be 0.0000001.
`maxit`	an (optional) integer value which represents the maximum number of iterations allowed for the fitting algorithm. By default, `maxit` is set to be 50.
`adjr2`	an (optional) logical variable. If TRUE, the adjusted deviance-based R-squared is calculated. By default, `adjr2` is set to be FALSE.
`...`	further arguments passed to or from other methods.

Details

If the value of waves for a cluster of size 4 is 2, 4, 5, 6 then it means that the data on times 1 and 3 are missing, which should be taken into account by glmgee to estimate the correlation matrix when its structure is assumed to be "Unstructured", "Stationary-M-dependent", "Non-Stationary-M-dependent" or "AR-1". If in this scenario waves is not specified then glmgee assumes that the available data for this cluster were taken on point times 1, 2, 3 and 4.

Value

an object of the class glmgee in which are stored the main results of the generalized estimating equation fitted to the data. Some of those results can be easily accessed using functions as, for example, print(), summary(), model.matrix(), estequa(), coef(), vcov(), logLik(), fitted(), confint() and predict(). In addition, the model fitted to the data can be assessed using functions as, for instance, anova.glmgee, residuals.glmgee, dfbeta.glmgee and cooks.distance.glmgee. The variable selection may be accomplished using stepCriterion.glmgee whereas the working–correlation–structure can be chosen by using criteria as QIC, CIC, DEW and RJC.

References

Liang K.Y. and Zeger S.L. (1986) Longitudinal data analysis using generalized linear models. Biometrika 73, 13-22.

Zeger S.L. and Liang K.Y. (1986) Longitudinal data analysis for discrete and continuous outcomes. Biometrics 42, 121-130.

Hardin J.W. and Hilbe J.M. (2013). Generalized Estimating Equations. Chapman & Hall, London.

Examples

## Example 1
mod1 <- size ~ poly(days,4) + treat
fit1 <- glmgee(mod1, id=tree, family=Gamma("log"), corstr="AR-1", data=spruce)
summary(fit1)

## Example 2
mod2 <- depressd ~ visit + group
fit2 <- glmgee(mod2, id=subj, family=binomial("logit"), corstr="Exchangeable", data=depression)
summary(fit2)

## Example 3
mod3 <- dep ~ visit + group
fit3 <- glmgee(mod3, id=subj, corstr="Non-Stationary-M-dependent(2)", data=depression)
summary(fit3)

glmtoolbox

Set of Tools to Data Analysis using Generalized Linear Models

v0.1.0

GPL-2 | GPL-3

Authors

Luis Hernando Vanegas [aut, cre], Luz Marina Rondón [aut], Gilberto A. Paula [aut]

Initial release