| NEWS | R Documentation |
Some adjustments to default parameters.
Use median, not mean, for the interval search for new points. Use Neumaier compensated addition in computation of gradients.
Some more documentation.
Fixed a bug in the interval censored gradient which slowed down convergence dramatically.
Fixed a bug which prevents calling mphcrm from within a
function with variables defined there. Like in an mclapply which
creates the data frame.
Fixed a bug which prevented the use of factors when running on a cluster.
Remove covariates and factor levels which are constant for a
transition. They will act as intercepts, we don't want that. If you
really want such a thing, create it with some neglible variation,
e.g. as rnorm(N,1,10*.Machine$double.eps).
Added generics for vcov(), coef().