library("countreg")
data("CrabSatellites", package = "countreg")
## default start values
zt_nb <- zerotrunc(satellites ~ width + as.numeric(color), data = CrabSatellites,
subset = satellites > 0, dist = "negbin")
## user-supplied start values and other options
zt_nb2 <- zerotrunc(satellites ~ width + as.numeric(color), data = CrabSatellites,
subset = satellites > 0, dist = "negbin", start = c(0.5, 0, 0))Control Parameters for Zero-Truncated Count Data Regression
Description
Various parameters that control fitting of zero-truncated count regression models using zerotrunc.
Usage
zerotrunc.control(method = "BFGS", maxit = 10000, start = NULL, ...)
Arguments
method
|
characters string specifying the method argument passed to optim.
|
maxit
|
integer specifying the maxit argument (maximal number of iterations) passed to optim.
|
start
|
an optional vector of starting values, see details. |
…
|
arguments passed to optim.
|
Details
All parameters in zerotrunc are estimated by maximum likelihood using optim with control options set in zerotrunc.control. Most arguments are passed on directly to optim, only start is used to control how optim is called.
Starting values can be supplied via start or estimated by glm.fit (default). Standard errors are derived numerically using the Hessian matrix returned by optim. To supply starting values, start should be a vector with (at least) starting values for the regression coefficients. In case a negative binomial distribution with unknown theta is used, a starting value for theta may be supplied by adding an additional vector element (e.g., start = c(coef, theta)); by default theta = 1 is used as the starting value otherwise.
Value
A list with the arguments specified.
See Also
zerotrunc