arf3DS4 (version 2.5-3)

processSeed: Process a model using a seed number of regions

Description

processSeed runs a sequence of models with different starting values to obtain a stable solution. It starts with as simple model with rectangular starts, subsequently it runs a full model given the estimates of the simple model. It then prunes the full model, discarding the regions with bad gradients and/or bounded parameters.

Usage

processSeed(modelname='defaultmodel', seedreg, subject='', 
 condition='', startmethod=c('default','simple'), grad=NULL, 
 bound=NULL, pval=NULL, options=new('options'), pr=T, 
 printlevel=0, try.silen=T, overwrite=T, experiment=NULL)

Arguments

modelname

Name to give the pruned models in the sequence.

seedreg

Number of regions to fit at sequence start (seed number).

subject

Name of the subject.

condition

Name of the condition.

startmethod

Use simple model as start or the default (rectangular method).

grad

Gradient check limits, absolute higher values are pruned.

bound

Threshold to check if a parameter is on a bound. When NULL, check is suppressed.

pval

Prune regions with non-significant Wald statistics higher than pval.

options

Options object.

pr

Output to screen?

printlevel

Number indicating the level of output of the minimization procedure.

try.silen

Should errors be printed when they occur or only passed to the warnings slot.

overwrite

Overwrite the model if a model with the same name already exists?

experiment

Optional experiment object.

Value

• Returns the best model (class model)

See Also

model, newModel, fitModel, pruneModel, options, experiment