[R] Parameters of Beta distribution

[David Winsemius]

On Oct 8, 2009, at 1:13 PM, Albyn Jones wrote:

> Quoting David Winsemius <dwinsemius at comcast.net>:
>
>
>> In insurance situation there is typically a cap on the covered  
>> losses and there is also typically an amount below which it would  
>> not make sense to offer a policy. So a minimum and a maximum are  
>> sensible assumptions about loss distributions in may real modeling  
>> situations.
>>
>> -- 
>> David.
>
> is that cap the same for every policy, or are there different types  
> of policies with different bounds?

A variation in policy coverage maximums is typical for casualty and  
life insurance, and the contingencies are of different types  
(variable, fixed). There is often a cap at $US 10^6 on US health  
insurance coverage (also variable). The actuarial problem for casualty  
coverage distributions is handled with some sort of convolution  
operation. (ObIANAA.)

> If there are caps, then other probability models mentioned like the  
> Weibull don't seem reasonable, unless they are truncated  
> distribution models.

Agreed. Hopefully that issue will be exposed in the process of  
comparing the fit of the distributions having appropriate structure to  
the problem to those having an inappropriate structure. I would also  
wonder how one can fairly compare across distribution types with  
varying numbers of parameters?

>
> albyn

David Winsemius, MD
Heritage Laboratories
West Hartford, CT