Skip to content

running the aggregate claims distribution with the package "actuar"

2 messages · Charles Thuo, Christophe Dutang

Original
Charles Thuo
# 
# i  run the following after fitting claims data into the negative binomial
as the frequency  distribution and the lognormal for the severity
distribution using the package "fitdistrplus".

require(actuar)


fx<- discretize(plnorm(x,11.69,0.79331),method="unbiased",step=500,from=0,


        to=qlnorm(1-1E-6,11.69,0.79331),lev=levlnorm(x,11.69,0.79331))



 pb<- 2.6/18.6


 pn<- dnbinom(0:qnbinom(1-1E-6,size=2.6,prob=pb),size=2.6,prob=pb)



 Fs<-
aggregateDist("convolution",model.sev=fx,model.freq=pn,x.scale=1,tol=1E-6)


# The Fs function stalls and does not produce and output. Can anyoen advise
why this is the case?

# On the other hand does the convolution method require iterations.
# In case the function Fs run successfully how are the parameters of the Fs
distribution extracted. How would the same parameters be related to an
existing portfolio of outstanding claims.
Christophe Dutang
# 
Hi,

Did you increase the maximum number of iterations? maxit argument

Regards, Christophe
?
Christophe Dutang
LMM, UdM, Le Mans, France
web: http://dutangc.free.fr

Le 23 oct. 2014 ? 07:20, Charles Thuo <tcmuigai at gmail.com> a ?crit :