The hardware and bandwidth for this mirror is donated by dogado GmbH, the Webhosting and Full Service-Cloud Provider. Check out our Wordpress Tutorial.
If you wish to report a bug, or if you are interested in having us mirror your free-software or open-source project, please feel free to contact us at mirror[@]dogado.de.
First release.
pcdreg() fits the semiparametric proportional rate
model for panel count data with time-varying covariates, using the EM
algorithm of Sun et al. (2024).
pcdreg(model = "mean") fits the proportional means
model by the estimating equation of Hu, Sun and Wei (2003). It is the
comparator used in the application of Sun et al. (2024), and is included
so that the two model families can be compared on the same data. Note
that the two are not reparametrisations of each other when covariates
vary over time, and that the fitted mean function is not constrained to
increase.
Covariance estimation for the rate model follows the paper
directly. vcov() supports type = "robust" (the
sandwich estimator, consistent without the Poisson assumption),
type = "information" (the efficient information estimator),
and type = "profile" (the Murphy–van der Vaart profile
likelihood estimator, computed only when requested via
pcdreg(profile = TRUE)). The means model has no likelihood,
so only the sandwich estimator is available for it.
Model specification uses pcd() on the left hand side
of the formula, following the Surv() convention:
pcd(id, time, count) for examination-time rows and
pcd(id, tstart, tstop, count) for counting-process rows
carrying time-varying covariates.
Both fitters return an object inheriting from
"pcdfit", so coef(), vcov(),
confint(), summary(), baseline(),
predict() and plot() work the same way for
either.
tidy(), glance() and
augment() methods are provided for the broom generics.
tidy() takes the same type argument as
vcov(), so the tidy path reports the same standard errors
as summary(). augment() adds the fitted mean,
the observed cumulative count and their difference to each row.
sim_pcd() simulates from the design of the paper’s
simulation study.
autoplot() methods return ggplot
objects for the data, drawn as one tile per examination interval shaded
by its count, and for the fitted baseline. plot() draws
them and returns the object invisibly.
baseline(), predict(type = "mean") and
sim_pcd() return tibbles.
Errors, warnings and messages use cli and carry condition
classes, so a particular failure can be caught with
tryCatch() rather than by matching on message
text.
These binaries (installable software) and packages are in development.
They may not be fully stable and should be used with caution. We make no claims about them.
Health stats visible at Monitor.