Hi Giorgio,

I run into a problem with the function mbbefdExposure. It doesn’t seem to like parameter of c>4.1, in particular I can’t reproduce the Lloyd’s curve (c=5):

mbbefdExposure(0.5, b=swissRe(5.0)['b'], g=swissRe(5.0)['g'])
b
NaN
Warning messages:
1: In log(a + b^x) : NaNs produced
2: In log(a + b) : NaNs produced

I ended up writing the following function, taken from the Bernegger paper:

mbbefdCurve <- function(x, C){
g <- exp((0.78 + 0.12_C)C)
b <- exp(3.1 - 0.15(1+C)_C)
if(identical(g, 1) | identical(b, 0))
x
if(identical(b, 1) & g > 1)
log(1+(g-1)_x)/log(g)
if(identical(b_g, 1) & g>1)
(1-b^x)/(1-b)
else
log(((g-1)_b + (1-g_b)_b^x)/(1-b))/log(g_b)
}

Hi, Cran says:
so I would move all the help into roxygen2 files.

Dear maintainers,

This concerns CRAN packages maintained by one of you, see the long list
at the end of this message.

These packages document package metadata (specifically, package
versions) in their *-package Rd file(s) which have gotten out of sync
with the actual package metadata in the DESCRIPTION file.

Most likely, these Rd files were generated by older versions of
utils::promptPackage() (maybe via utils::package.skeleton()), which
generated package documentation shells/skeletons with copies of the
current package metadata. This was not a good idea, as inevitably such
copies go out of sync with their master values. The R 3.2.0 release
(3 years ago) did

A number of macros have been added in the new 'share/Rd' directory for
use in package overview help pages, and promptPackage() now makes use
of them.

(so that package documentation shells/skeletons generated with current
versions of utils::promptPackage() use Rd macro references to the
package metadata values instead of making copies).

Can you please bring package Rd files "up to date"? The simplest may be
using a current utils::promptPackage() to re-generate the
shell/skeleton, and merge with the content of the current package Rd
file. (In case the updated package Rd file would more or less only
duplicate the package metadata, feel free to drop the package Rd file
from your package sources.)

Best
-k

Maybe it has been already solved btw I copy and paste

Dear maintainer,

Presumably following a recent update of 'fitdistrplus', we now see

• checking R code for possible problems ... NOTE
  fitDR: warning in mledist(x, distr = "mbbefd", start =
  initparmbbefd[[1]], custom.optim = constrOptim.nl, hin =
  constrmbbefd1, control.outer = list(trace = FALSE), gr = grLL):
  partial argument match of ‘gr’ to ‘gradient’
  fitDR: warning in mledist(x, distr = "mbbefd", start =
  initparmbbefd[[2]], custom.optim = constrOptim.nl, hin =
  constrmbbefd2, control.outer = list(trace = FALSE), gr = grLL):
  partial argument match of ‘gr’ to ‘gradient’
• checking examples ... ERROR
  Running examples in ‘mbbefd-Ex.R’ failed
  The error most likely occurred in:

base::assign(".ptime", proc.time(), pos = "CheckExEnv")

Name: fitDR

Title: Fit of destruction rate models

Aliases: fitDR

Keywords: distribution

** Examples

(1) fit of a one-inflated beta distribution by maximum likelihood estimation

n <- 1e3
set.seed(12345)
x <- roibeta(n, 3, 2, 1/6)

f1 <- fitDR(x, "oibeta", method="mle")
Warning in cbind(as.matrix(f1$vcov), rep(0, npar - 1)) :
number of rows of result is not a multiple of vector length (arg 2)
Warning in rbind(cbind(as.matrix(f1$vcov), rep(0, npar - 1)), c(rep(0, npar - :
number of columns of result is not a multiple of vector length (arg 2)
Error in fitDR(x, "oibeta", method = "mle") :
length of 'dimnames' [1] not equal to array extent
Execution halted

• checking for unstated dependencies in ‘tests’ ... OK
• checking tests ... ERROR
  Running the tests in ‘tests/test-beaonre.R’ failed.
  Last 13 lines of output:
  [1] "oistpareto"
  [1] "oibeta"
  Error in fitDR(x, d, method = "mle") :
  length of 'dimnames' [1] not equal to array extent
  Calls: lapply -> FUN -> fitDR
  In addition: Warning messages:
  1: In mledist(data, distname, start, fix.arg, ...) :
  The BFGS method cannot be used with bounds without provided the gradient. The method is changed to L-BFGS-B.
  2: In cov2cor(f1$vcov) :
  diag(.) had 0 or NA entries; non-finite result is doubtful
  3: In cbind(as.matrix(f1$vcov), rep(0, npar - 1)) :
  number of rows of result is not a multiple of vector length (arg 2)
  4: In rbind(cbind(as.matrix(f1$vcov), rep(0, npar - 1)), c(rep(0, npar - :
  number of columns of result is not a multiple of vector length (arg 2)
  Execution halted
• checking for unstated dependencies in vignettes ... OK
• checking package vignettes in ‘inst/doc’ ... OK
• checking re-building of vignette outputs ... WARNING
  Error in re-building vignettes:
  ...
  Loading required package: fitdistrplus
  Loading required package: MASS
  Loading required package: survival
  Loading required package: alabama
  Loading required package: numDeriv
  Package: mbbefd
  Version: 0.8-0
  Date: 2016-02-23 22:30:54
  BugReport: http://github.com/spedygiorgio/mbbefd/issues

Warning in cbind(as.matrix(f1$vcov), rep(0, npar - 1)) :
number of rows of result is not a multiple of vector length (arg 2)
Warning in rbind(cbind(as.matrix(f1$vcov), rep(0, npar - 1)), c(rep(0, npar - :
number of columns of result is not a multiple of vector length (arg 2)
Quitting from lines 682-688 (Introduction_to_mbbefd.Rmd)
Error: processing vignette ‘Introduction_to_mbbefd.Rmd’ failed with diagnostics:
length of ‘dimnames’ [1] not equal to array extent
Execution halted

Hello

Thank you for the great package.

I think the formula for G(x) in 2.1.2 within https://cran.r-project.org/web/packages/mbbefd/vignettes/Introduction_to_mbbefd.pdf isn’t correct.

The correct formula, when reduced, is I think B(x,a+1,b)+x*(a+b)/a*(1-B(x,a,))

Hi all,
it could be nice to add ORCID links if you have one R-Bloggers on Orcid

Forthcoming release of 3.4 R has changed something in compiled code management. See this GitHub issue GitHub Rcpp