spaMM (version 2.4.35)

update.HLfit: Updates an HLCor or HLfit fit

Description

update will update and (by default) re-fit a model. It does this mostly by extracting the call stored in the object, updating the call and evaluating that call. (however, currently the predictor argument is processed differently). Using update is a risky programming style (see Note).

Usage

# S3 method for HLfit
update(object, formula., ..., evaluate = TRUE)

Arguments

object

A return object from an HLfit call.

formula.

Changes to the formula. Beware of the syntax: see update.formula for details.

…

Additional arguments to the call, or arguments with changed values. Use name = NULL to remove the argument name.

evaluate

If TRUE, evaluate the new call else return the call.

Value

An HLCor or HLfit fit of the same type as the input object.

See Also

See also HLCor, HLfit.

Examples

# NOT RUN {
data("wafers")
## First the fit to be updated:
wFit <- HLfit(y ~X1*X3+X2*X3+I(X2^2)+(1|batch),family=Gamma(log),
          resid.model = ~ X3+I(X3^2) ,data=wafers)

# For estimates given by Lee et al., Appl. Stochastic Models Bus. Ind. (2011) 27:  315-328:
# Refit with given beta or/and phi values:
 
betavals <- c(5.55,0.08,-0.14,-0.21,-0.08,-0.09,-0.09)
# reconstruct fitted phi value from predictor for log(phi)
Xphi <- with(wafers,cbind(1,X3,X3^2)) ## design matrix
phifit <- exp(Xphi %*% c(-2.90,0.1,0.95))
update(wFit,formula.= . ~ offset(wFit$`X.pv` %*% betavals)+(1|batch),
       ranFix=list(lambda=exp(-3.67),phi=phifit))

## There are subtlety in performing REML fits of constrained models,
##   illustrated by thefact thatthe following fir does not recover 
##   the original likelihood values, because dispersion parameters are
##   estimated but the REML correction changes with the formula:
update(wFit,formula.= . ~ offset(wFit$`X.pv` %*% fixef(wFit))+(1|batch))
## To maintain the original REML correction, Consider instead
update(wFit,formula.= . ~ offset(wFit$`X.pv` %*% fixef(wFit))+(1|batch),
       REMLformula=wFit$predictor)  ## recover original p_v and p_bv 
## Alternatively, show original wFit as differences from betavals:  
update(wFit,formula.= . ~ . +offset(wFit$`X.pv` %*% betavals))

# }