Coerce a gev.fit Object into a NSGEV Object

Coerce a gev.fit object from the ismev package into a NSGEV object.

# S3 method for gev.fit
as.NSGEV(x, data, ...)

Arguments

x

The object to be coerced.

data

The matrix of covariates that was used.

...

Not used yet.

Value

An object with class "NSGEV".

Note

A gev.fit object does not embed the set of covariates used nor a even a call. Thus the set of covariates must be given. To be compliant with NSGEV, this data set must be a data frame or a matrix with colnames.

Author

Yves Deville

Examples

require(ismev)
## number of observations and covariates
n <- 20; m <- 4 

## generate a matrix covariates
set.seed(1234)
dat <- matrix(runif(n * m), nrow = n)
colnames(dat) <- rNames(m)

## response
y <- drop(nieve::rGEV(n))

## fit 
fit <- gev.fit(xdat = y,             ## response     
               ydat = dat,           ## matrix of covariates
               mul = 1:2, sigl = 4,  ## indices of covariates
               shl = NULL,           
               mulink = exp)         ## optional inverse-link
#> Warning: NaNs produced
#> $model
#> $model[[1]]
#> [1] 1 2
#> 
#> $model[[2]]
#> [1] 4
#> 
#> $model[[3]]
#> NULL
#> 
#> 
#> $link
#> [1] "c(exp, identity, identity)"
#> 
#> $conv
#> [1] 0
#> 
#> $nllh
#> [1] 30.22914
#> 
#> $mle
#> [1] -13.0806888   3.6753455  -6.1078553   1.2954442  -0.5832947   0.4902082
#> 
#> $se
#> [1]       NaN       NaN       NaN 0.3880493 0.5333207 0.5098481
#> 

## coerce
ns <- as.NSGEV(fit, data = dat)