R: Convert a smooth to a form suitable for estimating as random...

smooth2random {mgcv}

R Documentation

Convert a smooth to a form suitable for estimating as random effect

Description

A generic function for converting mgcv smooth objects to forms suitable for estimation as random effects by e.g. lme. Exported mostly for use by other package developers.

Usage

smooth2random(object,vnames,type=1)

Arguments

`object`	an `mgcv` smooth object.
`vnames`	a vector of names to avoid as dummy variable names in the random effects form.
`type`	`1` for `lme`, otherwise `lmer`.

Details

There is a duality between smooths and random effects which means that smooths can be estimated using mixed modelling software. This function converts standard mgcv smooth objects to forms suitable for estimation by lme, for example. A service routine for gamm exported for use by package developers.

Value

A list.

`rand`	a list of random effects, including grouping factors, and a fixed effects matrix. Grouping factors, model matrix and model matrix name attached as attributes, to each element.
`trans.D`	A vector, trans.D, that transforms coefs, in order [rand1, rand2,... fix] back to original parameterization. If null, then taken as vector of ones. `b.original = trans.U %% (trans.Db.fit)`.
`trans.U`	A matrix, trans.U, that transforms coefs, in order [rand1, rand2,... fix] back to original parameterization. If null, then not needed. If null then taken as identity.
`Xf`	A matrix for the fixed effects, if any.
`fixed`	`TRUE/FALSE`, indicating if term was unpenalized or not. If unpenalized then other stuff may not be returned (it's not a random effect).
`rind`	an index vector such that if br is the vector of random coefficients for the term, br[rind] is the coefs in order for this term.
`pen.ind`	index of which penalty penalizes each coefficient: 0 for unpenalized.

Author(s)

Simon N. Wood simon.wood@r-project.org.

References

Wood S.N. (2017) Generalized Additive Models: An Introduction with R (2nd edition). Chapman and Hall/CRC Press.

Examples

library(mgcv)
x <- runif(30)
sm <- smoothCon(s(x),data.frame(x=x))[[1]]
smooth2random(sm,"")

[Package mgcv version 1.8-23 Index]