dynfrail_fit: Inner maximization of the log-likelihood
In dynfrail: Fitting Dynamic Frailty Models with the EM Algorithm

Description Usage Arguments Details Value Examples

Inner maximization of the log-likelihood

1 2	dynfrail_fit(logfrailtypar, dist, pvfm, Y, Xmat, atrisk, basehaz_line, mcox = list(), c_vecs, inner_control, return_loglik = TRUE)

`logfrailtypar`	A vector containing the natural logarithm of the two parameters (`theta` - for the distribution, `lambda` - for the autocorrelation)
`dist`	Argument of `dynfrail_dist`
`pvfm`	Argument of `dynfrail_dist`
`Y`	A `Surv` object obtained by splitting the original data at all the time points where the frailty process takes new values
`Xmat`	A model matrix obtained by splitting the original data at all the time points where the frailty process takes new values
`atrisk`	A list of various calculations that are used in the maximization process.
`basehaz_line`	A vector with the baseline hazard estimate at each right hand side time point from `Y` (can be 0 for the others)
`mcox`	An initial Cox model
`c_vecs`	A list of the length equal to the number of clusters; each element contains a vector of the length of different values that Z(t) takes in that cluster. Each element of this vector contains the sum of the cumulative hazards associated with that value of the frailty.
`inner_control`	Argument of `dynfrail_control`
`return_loglik`	Logical. If `TRUE`, then this just returns the log-likelihood, otherwise it returns also the estimates and information matrix

This is an internal function that is used by dynfrail. It is not recommended to use this directly unless you know exactly what you are doing. On the other hand, this might be useful if someone wants, for example, to use different maximizers, or to calculate the log-likelihood at specific values of theta, lambda. Most of the input can be produced by dynfrail_prep.

The log-likelihood if return_loglik = TRUE or a list with the log-likelihood and estimates if return_loglik = FALSE.

arglist1 <- dynfrail_prep(Surv(time, status) ~ rx + sex + cluster(litter),
data = rats)

# using list() inside is because of the way that R converts lists and vectors
mod1 <- do.call(dynfrail_fit, c(logfrailtypar = list(log(c(0.5, 0.1))), arglist1))