pfdaControl: Control Setting for pfda
In pfda: Paired Functional Data Analysis

Description Usage Arguments Details Value Author(s) See Also Examples

This gives the settings and defaults for fitting the EM algorithms in the pfda function

pfdaControl(..., penalty.method = c("AIC", "CV"), minimum.variance = 1e-04, convergence.tolerance = 0.01, max.iterations = 10000, nfolds = 10,  binary.k0 = 100, binary.kr = 10, binary.burnin = 100, nknots=11)

`...`	any extra arguments to be passed in.
`penalty.method`	the method for optimizing penalties. Either CV for cross validation or AIC for Akaike's an information criteria.
`minimum.variance`	The minimum allowable as a variance.
`convergence.tolerance`	The tolerance for determining convergence of the EM algorithm.
`max.iterations`	The maximum number of Iterations before determining the algorithm failed.
`nfolds`	Number of folds in cross validation.
`binary.k0`	The number of simulated runs for the initial steps in the stocastic approximation involved with the binary drivers.
`binary.kr`	the ongoing number of simulation runs for the stocastic approximation for binary drivers.
`binary.burnin`	The length of the burn in period for stochasitc approximation.
`nknots`	the number of internal knots for the splines. Ignored if knots are specified.

This is a convenience function for specifying a list of control parameters that control the fit of the EM aglorithm. With the ... argument there are several arguments that can be additionally specified. optim.method controls the method used in optim to optimize the penalties. The optim.start can be specified to give a starting value for the optimization of the penalties.

a list with the class set to pfdaControl. Named elements are the same as the named arguments, plus any additional named arguments given in ...

Andrew Redd

pfda