mphcrm stores coefficients in a list,
not in a vector. This is because they should be treated
differently according to whether they are probabilities,
proportional hazards, or coefficients for factor levels or
ordinary covariates. flatten extracts them as a named
vector. unflatten puts them back in structured form.
flatten(x, exclude = attr(x, "exclude"))unflatten(
flesh,
skeleton = attr(flesh, "skeleton"),
exclude = attr(flesh, "exclude")
)
parameter set as typically found in opt[[1]]\$par,
where opt is returned from mphcrm.
For internal use
vector of class "relistable", as returned from flatten.
For internal use
flatten/unflatten is just a thinly disguised
unlist/relist, but uses
slightly more readable names.