The R Project SVN R-packages

\name{bam.update}
\alias{bam.update}

%- Also NEED an `\alias' for EACH other topic documented here.
\title{Update a strictly additive bam model for new data.}

\description{ Gaussian with identity link models fitted by \code{\link{bam}} can be efficiently updated as new data becomes available,
by simply updating the QR decomposition on which estimation is based, and re-optimizing the smoothing parameters, starting
from the previous estimates. This routine implements this.
}
\usage{
bam.update(b,data,chunk.size=10000)
}
%- maybe also `usage' for other objects documented here.

\arguments{ 
\item{b}{ A \code{gam} object fitted by \code{\link{bam}} and representing a strictly additive model 
(i.e. \code{gaussian} errors, \code{identity} link).}
\item{data}{Extra data to augment the original data used to obtain \code{b}. Must include a \code{weights} column if the 
            original fit was weighted and a \code{AR.start} column if \code{AR.start} was non \code{NULL} in original fit.}
\item{chunk.size}{size of subsets of data to process in one go when getting fitted values.}
}

\value{ 
An object of class \code{"gam"} as described in \code{\link{gamObject}}.
}


\details{ \code{bam.update} updates the QR decomposition of the (weighted) model matrix of the GAM represented by \code{b} to take 
account of the new data. The orthogonal factor multiplied by the response vector is also updated. Given these updates the model 
and smoothing parameters can be re-estimated, as if the whole dataset (original and the new data) had been fitted in one go. The 
function will use the same AR1 model for the residuals as that employed in the original model fit (see \code{rho} parameter 
of \code{\link{bam}}).

Note that there may be small numerical differences in fit between fitting the data all at once, and fitting in 
stages by updating, if the smoothing bases used have any of their details set with reference 
to the data (e.g. default knot locations).

}


\author{ Simon N. Wood \email{simon.wood@r-project.org}


}
\section{WARNINGS }{
AIC computation does not currently take account of AR model, if used.
}

\seealso{\code{\link{mgcv-package}}, \code{\link{bam}}
 }

\examples{
library(mgcv)
## following is not *very* large, for obvious reasons...
set.seed(8)
n <- 5000
dat <- gamSim(1,n=n,dist="normal",scale=5)
dat[c(50,13,3000,3005,3100),]<- NA
dat1 <- dat[(n-999):n,]
dat0 <- dat[1:(n-1000),]
bs <- "ps";k <- 20
method <- "GCV.Cp"
b <- bam(y ~ s(x0,bs=bs,k=k)+s(x1,bs=bs,k=k)+s(x2,bs=bs,k=k)+
           s(x3,bs=bs,k=k),data=dat0,method=method)

b1 <- bam.update(b,dat1)

b2 <- bam.update(bam.update(b,dat1[1:500,]),dat1[501:1000,])
 
b3 <- bam(y ~ s(x0,bs=bs,k=k)+s(x1,bs=bs,k=k)+s(x2,bs=bs,k=k)+
           s(x3,bs=bs,k=k),data=dat,method=method)
b1;b2;b3

## example with AR1 errors...

e <- rnorm(n)
for (i in 2:n) e[i] <- e[i-1]*.7 + e[i]
dat$y <- dat$f + e*3
dat[c(50,13,3000,3005,3100),]<- NA
dat1 <- dat[(n-999):n,]
dat0 <- dat[1:(n-1000),]

b <- bam(y ~ s(x0,bs=bs,k=k)+s(x1,bs=bs,k=k)+s(x2,bs=bs,k=k)+
           s(x3,bs=bs,k=k),data=dat0,rho=0.7)

b1 <- bam.update(b,dat1)


summary(b1);summary(b2);summary(b3)

}

\keyword{models} \keyword{smooth} \keyword{regression}%-- one or more ..
\concept{Varying coefficient model}
\concept{Functional linear model}
\concept{Penalized GLM}
\concept{Generalized Additive Model}
\concept{Penalized regression}
\concept{Spline smoothing}
\concept{Penalized regression spline}
\concept{Generalized Cross Validation}
\concept{Smoothing parameter selection}
\concept{tensor product smoothing}
\concept{thin plate spline}
\concept{P-spline}
\concept{Generalized ridge regression}