refund (version 0.1-13)

amc: Additive model with constraints

Description

An internal function, called by fosr(), that fits additive models with linear constraints via a call to gam or bam in the mgcv package.

Usage

amc(y, Xmat, S, gam.method = "REML", C = NULL, lambda = NULL, ...)

Arguments

y

response vector.

Xmat

design matrix.

S

list of penalty matrices.

gam.method

smoothing parameter selection method: "REML" for restricted maximum likelihood, "GCV.Cp" for generalized cross-validation.

C

matrix of linear constraints. Dimension should be number of constraints times ncol(Xmat).

lambda

smoothing parameter value. If NULL, the smoothing parameter(s) will be estimated.

...

other arguments, passed to gam or bam.

Value

• A list with the following elements:
• gam
  the gam object returned by gam or bam.
• coefficients
  coefficients with respect to design matrix Xmat, derived from the gam() fit.
• Vp, GinvXt
  outputs used by fosr.
• method
  the gam.method argument of the call to amc.

Details

The additive model is fitted using gam, unless there are more than 10000 responses; in that case bam is used.

See Also

fosr