`"twinstim"`

Objects`twinstim_methods.Rd`

Besides `print`

and `summary`

methods there
are also some standard extraction methods defined for objects of class
`"twinstim"`

: `vcov`

, `logLik`

, and
`nobs`

. This
also enables the use of, e.g., `confint`

and
`AIC`

. The model `summary`

can be exported to LaTeX
by the corresponding `toLatex`

or `xtable`

methods.

```
# S3 method for twinstim
print(x, digits = max(3, getOption("digits") - 3), ...)
# S3 method for twinstim
summary(object, test.iaf = FALSE,
correlation = FALSE, symbolic.cor = FALSE, runtime = FALSE, ...)
# S3 method for twinstim
coeflist(x, ...)
# S3 method for twinstim
vcov(object, ...)
# S3 method for twinstim
logLik(object, ...)
# S3 method for twinstim
nobs(object, ...)
# S3 method for summary.twinstim
print(x,
digits = max(3, getOption("digits") - 3), symbolic.cor = x$symbolic.cor,
signif.stars = getOption("show.signif.stars"), ...)
# S3 method for summary.twinstim
toLatex(object,
digits = max(3, getOption("digits") - 3), eps.Pvalue = 1e-4,
align = "lrrrr", booktabs = getOption("xtable.booktabs", FALSE),
withAIC = FALSE, ...)
# S3 method for summary.twinstim
xtable(x, caption = NULL, label = NULL,
align = c("l", "r", "r", "r"), digits = 3,
display = c("s", "f", "s", "s"), ...,
ci.level = 0.95, ci.fmt = "%4.2f", ci.to = "--",
eps.Pvalue = 1e-4)
```

- x, object
an object of class

`"twinstim"`

or`"summary.twinstim"`

, respectively.- digits
integer, used for number formatting with

`signif()`

. Minimum number of significant digits to be printed in values.- test.iaf
logical indicating if the simple Wald z- and p-values should be calculated for parameters of the interaction functions

`siaf`

and`tiaf`

. Because it is often invalid or meaningless to do so, the default is`FALSE`

.- correlation
logical. If

`TRUE`

, the correlation matrix of the estimated parameters is returned and printed.- symbolic.cor
logical. If

`TRUE`

, print the correlations in a symbolic form (see`symnum`

) rather than as numbers.- runtime
logical. If

`TRUE`

, the summary additionally includes the time elapsed and the number of log-likelihood and score function evaluations during model fitting.- signif.stars
logical. If

`TRUE`

, “significance stars” are printed for each coefficient.- eps.Pvalue
passed to

`format.pval`

.- booktabs
logical indicating if the

`toprule`

,`midrule`

and`bottomrule`

commands from the LaTeX package booktabs should be used for horizontal lines rather than`hline`

.- withAIC
logical indicating if the AIC and the log-likelihood of the model should be included below the table of coefficients in the LaTeX tabular.

- caption,label,align,display
see

`xtable`

.- ci.level,ci.fmt,ci.to
the confidence intervals are calculated at level

`ci.level`

and printed using`sprintf`

with format`ci.fmt`

and separator`ci.to`

.- ...
For

`print.summary.twinstim`

, arguments passed to`printCoefmat`

.

For all other methods: unused (argument of the generic).

The estimated coefficients and standard Wald-type confidence intervals
can be extracted using the default `coef`

and
`confint`

methods from package stats.
Note, however, that there is the useful `coeflist`

method to
list the coefficients by model component.

The `print`

and `summary`

methods allow the compact or comprehensive
representation of the fitting results, respectively. The former only prints
the original function call, the estimated coefficients and the maximum
log-likelihood value. The latter prints the whole coefficient matrix
with standard errors, z- and p-values (see `printCoefmat`

)
-- separately for the endemic and the epidemic component -- and
additionally the AIC, the achieved log-likelihood, the number of
log-likelihood and score evaluations, and the runtime.
They both append a big “WARNING”, if the optimization algorithm
did not converge.

The `toLatex`

method is essentially a
translation of the printed summary table of coefficients to LaTeX
code (using xtable). However, the `xtable`

method does a
different job in that it first converts coefficients to rate ratios
(RR, i.e., the `exp`

-transformation) and gives confidence
intervals for those instead of standard errors and z-values.
Intercepts and interaction function parameters are ignored by the
`xtable`

method.

The `print`

methods return their first argument, invisibly, as
they always should.
The `vcov`

method returns the estimated variance-covariance
matrix of the parameters, which is the inverse of
`object$fisherinfo`

(estimate of the *expected* Fisher
information matrix). This `"fisherinfo"`

is not always available
(see `twinstim`

), in which case
`object$fisherinfo.observed`

is used if available or an error is
returned otherwise.
The `logLik`

and `nobs`

methods return the maximum
log-likelihood value of the model, and the number of events (excluding
events of the prehistory), respectively.

The `summary`

method returns a list containing some summary
statistics of the model, which is nicely printed by the corresponding
`print`

method.

The `toLatex`

method returns a character vector of class
`"Latex"`

, each element containing one line of LaTeX code (see
`print.Latex`

).
The `xtable`

method returns an object of class
`"xtable"`

. Note that the column name of the confidence
interval, e.g. “95% CI”, contains the percent symbol that may
need to be escaped when printing the `"xtable"`

in the output
format (see `sanitize.text.function`

in
`print.xtable`

). This may also hold for row names.

Sebastian Meyer

```
# load a fit of the 'imdepi' data, see the example in ?twinstim
data("imdepifit")
# print method
imdepifit
# extract point estimates (in a single vector or listed by model component)
coef(imdepifit)
coeflist(imdepifit)
# variance-covariance matrix of endemic parameters
# (inverse of expected Fisher information)
unname(vcov(imdepifit)[1:4,1:4])
# the default confint() method may be used for Wald CI's
confint(imdepifit, parm="e.typeC", level=0.95)
# log-likelihood and AIC of the fitted model
logLik(imdepifit)
AIC(imdepifit)
nobs(imdepifit)
# produce a summary with parameter correlations and runtime information
(s <- summary(imdepifit, correlation=TRUE, symbolic.cor=TRUE, runtime=TRUE))
# create LaTeX code of coefficient table
toLatex(s, withAIC=FALSE)
# or using the xtable-method (which produces rate ratios)
xtable(s)
```