mphcrm.control: Control parameters for mphcrm
In durmod: Mixed Proportional Hazard Competing Risk Model

Modify the default estimation parameters for mphcrm.

1	mphcrm.control(...)

...

parameters that can be adjusted. See the vignette("whatmph") for more details.

threads. integer. The number of threads to use. Defaults to getOption('durmod.threads').
iters. integer. How many iterations should we maximally run. Defaults to 50.
ll.improve. numeric. How much must the log-likelihood improve from the last iteration before termination. Defaults to 0.001.
newpoint.maxtime. numeric. For how many seconds should a global search for a new point improving the likelihood be conducted before we continue with the best we have found. Defaults to 120.
callback. A user-specified function(fromwhere, opt, dataset, control, ...) which is called after each optimization step. It can be used to report what is happening, check whatever it wants, and optionally stop the estimation by calling stop(). In this case, mphcrm() will return with the most recently estimated set of parameters. See the help on mphcrm.callback for information on the argument.
trap.interrupt. logical. Should interrupts be trapped so that mphcrm returns gracefully? In this case the program will continue. Defaults to interactive().
cluster. Cluster specification from package parallel or snow.

List of control parameters suitable for the control argument of mphcrm.

There are more parameters documented in the vignette("whatmph"). Some of them can be useful. Instead of cluttering the source code with constants and stuff required by various optimization routines, they have been put in this control list.

durmod documentation built on March 31, 2020, 5:23 p.m.