bGAMMControl {GMMBoost}

Control Values for bGAMM fit

Description

The values supplied in the function call replace the defaults and a list with all possible arguments is returned. The returned list is used as the control argument to the bGAMM function.

Usage

bGAMMControl(nue=0.1,add.fix=NULL,start=NULL,q_start=NULL, 
             OPT=TRUE,nbasis=20,spline.degree=3,
             diff.ord=2,sel.method="aic",steps=500,
             method="EM",overdispersion=FALSE)  

Arguments

nue

weakness of the learner. Choose 0 < nue =< 1. Default is 0.1.

add.fix

a vector specifying smooth terms, which are excluded from selection.

start

a vector containing starting values for fixed and random effects of suitable length. Default is a vector full of zeros.

q_start

a scalar or matrix of suitable dimension, specifying starting values for the random-effects variance-covariance matrix. Default is a scalar 0.1 or diagonal matrix with 0.1 in the diagonal.

OPT

logical scalar. When TRUE the estimates at the optimal number of boosting steps, chosen by information criteria, are derived. If FALSE, the estimates at the maximal number of boosting steps are derived. Default is TRUE.

nbasis

the number of b-spline basis functions for the modeling of smooth terms. Default is 20.

spline.degree

the degree of the B-spline polynomials. Default is 3.

diff.ord

the order of the difference penalty; must be lower than the degree of the B-spline polynomials (see previous argument). Default is 2.

sel.method

two different information criteria, "aic" or "bic", can be chosen, on which the selection step is based on. Default is "aic".

steps

the number of boosting interations. Default is 500.

method

two methods for the computation of the random-effects variance-covariance parameter estimates can be chosen, an EM-type estimate and an REML-type estimate. The REML-type estimate uses the bobyqa function for optimization. Default is EM.

overdispersion

logical scalar. If FALSE, no scale parameter is derived, if TRUE, in each boosting iteration a scale parameter is estimated by use of Pearson residuals. This can be used to fit overdispersed Poisson models. Default is FALSE.

Value

a list with components for each of the possible arguments.

Author(s)

Andreas Groll [e-mail]

See Also

bGAMM, bobyqa

Examples

# decrease the maximum number of boosting iterations 
# and use BIC for selection
bGAMMControl(steps = 100, sel.method = "BIC")