Dear SAS-L,

I am currently trying to get a grip on so-called IBNR (incurred but
not resolved) estimation of health care insurance claims, based on
historical patterns. I already programmed (rather copied) an approach
by Mack based on the Chain Ladder technique. I used arrays in a
datastep program, which runs fast, but will become unreadable very
soon.

I am wondering if there is any experience, sample code, macro, IML,
anything, that would help me on this. I am thinking about AI
techniques (genetic algorithms?) to find a minimum variance estimator,
perhaps bootstrapping techniques to estimate the confidence interval
and so on.

Googleing on the subject (SAS and IBNR) gives very little hits.

Hope you can help me on this!

Thanks and Happy Holidays,
Eric

Eric,

I don't know what it includes, or its cost, but SAS offers a health
insurance package.

Alternatively, I saw the following that is available for R:
http://actuaryzhang.com/software//software.html#ChainLadder

HTH,
Art
------------
On Dec 24, 5:00=A0am, Eric Hoogenboom <erichoogenb...@yahoo.com> wrote:
> Dear SAS-L,
>
> I am currently trying to get a grip on so-called IBNR (incurred but
> not resolved) estimation of health care insurance claims, based on
> historical patterns. I already programmed (rather copied) an approach
> by Mack based on the Chain Ladder technique. I used arrays in a
> datastep program, which runs fast, but will become unreadable very
> soon.
>
> I am wondering if there is any experience, sample code, macro, IML,
> anything, that would help me on this. I am thinking about AI
> techniques (genetic algorithms?) to find a minimum variance estimator,
> perhaps bootstrapping techniques to estimate the confidence interval
> and so on.
>
> Googleing on the subject (SAS and IBNR) gives very little hits.
>
> Hope you can help me on this!
>
> Thanks and Happy Holidays,
> Eric

someone has developed something similar? 

I'm developing now a Chain Ladder procedure for IBNR, but it is some complicated