stripHLfit: Reduce the size of fitted objects

In spaMM: Mixed-Effect Models, with or without Spatial Random Effects

Reduce the size of fitted objects

Description

Large matrices and other memory-expensive objects may be stored in a fit object. This function removes them in order to reduce the size of the object, particularly when stored on disk. In principle, the removed objects can be regenerated automatically when needed (e.g., for a predict()).

Usage

stripHLfit(object, ...)

Arguments

object	The result of a fit (an object of class HLfit).
...	Further arguments, not currently used.

Value

The input fit objects with some elements removed.

Note

The effect may change without notice between versions as the efficiency of the operation is highly sensitive to implementation details.

Examples

## Not run: 
## rather unconvincing example : quantitative effect is small.

# measure size of saved object:
saveSize <- function (object,...) {
    tf <- tempfile(fileext = ".RData")
    on.exit(unlink(tf))
    save(object, file = tf,...)
    file.size(tf)
  }
data("Loaloa")  
lfit <- fitme(cbind(npos,ntot-npos)~elev1+elev2+elev3+elev4+maxNDVI1+seNDVI
                   +Matern(1|longitude+latitude), method="HL(0,1)",
              data=Loaloa, family=binomial(), fixed=list(nu=0.5,rho=1,lambda=0.5))
saveSize(lfit)                 
pfit <- predict(lfit,newdata=Loaloa,variances=list(cov=TRUE)) # increases size!
saveSize(lfit)
lfit <- stripHLfit(lfit)
saveSize(lfit)

## End(Not run)

spaMM documentation built on Aug. 30, 2023, 1:07 a.m.