smooth.construct.pss.smooth.spec.Rd

% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/pffr-utilities.R
\name{smooth.construct.pss.smooth.spec}
\alias{smooth.construct.pss.smooth.spec}
\title{P-spline constructor with modified 'shrinkage' penalty}
\usage{
\method{smooth.construct}{pss.smooth.spec}(object, data, knots)
}
\arguments{
\item{object}{see \code{\link[mgcv]{smooth.construct}}. The shrinkage factor can be specified via \code{object$xt$shrink}}

\item{data}{see \code{\link[mgcv]{smooth.construct}}.}

\item{knots}{see \code{\link[mgcv]{smooth.construct}}.}
}
\description{
Construct a B-spline basis with a modified difference penalty
of full rank (i.e., that also penalizes low-order polynomials).
}
\details{
This penalty-basis combination is useful to avoid non-identifiability issues for \code{\link{ff}} terms.
See 'ts' or 'cs' in \code{\link[mgcv]{smooth.terms}}
for similar "shrinkage penalties" for thin plate and cubic regression splines.
The basic idea is to replace the k-th zero eigenvalue of the original penalty by
\eqn{s^k \nu_m}, where \eqn{s} is the shrinkage factor (defaults to 0.1)
and \eqn{\nu_m} is the smallest non-zero eigenvalue. See reference for the
original idea, implementation follows that in the 'ts' and 'cs' constructors
(see \code{\link[mgcv]{smooth.terms}}).
}
\references{
Marra, G., & Wood, S. N. (2011). Practical variable selection for generalized additive models.
\emph{Computational Statistics & Data Analysis}, 55(7), 2372-2387.
}
\author{
Fabian Scheipl;  adapted from 'ts' and 'cs' constructors by S.N. Wood.
}