Re: Inverse-Gamma

CHEUNG Hung Francis ( h9607424@hkusua.hku.hk )
Sat, 27 Mar 1999 09:25:36 +0800

--=====================_2340917==_.ALT
Content-Type: text/plain; charset="us-ascii"

I have not touched the 4B stuff for almost a month (though i will take it in
May also!!!) I just want to tell what i can recall.
Regarding the primes, I think they are used so that you can distinguish between
parameters in the prior (y', m') and the current information (m periods of
observation, a total number of mE[ni] claims, and so an aggregate claim of y
(please refer to eqt 9.3 in Herzog))
a final word: if i were you, i would focus more on the meaning of parameters
instead of how they look like. "meaning" can enhance your understanding. ^_^

Francis

At 07:31 AM 3/27/99 , Gregor.Brown@libertynorthwest.com wrote:
>I have a question about the parameters for the Inverse-Gamma Distribution.
>Can anyone humor me as to why there are primes ( ' ) on the parameters. Is
>there any good reason or is it just done so actuarial students can use even
>more symbol notation?
>shape parameter: m'
>scale: y'
>
>mean = y'/(m'-2)
>etc...
>
>Thanks,
>Gregor L. Brown
>

--=====================_2340917==_.ALT
Content-Type: text/html; charset="us-ascii"

I have not touched the 4B stuff for almost a month (though iwill take it in May also!!!) I just want to tell what i can recall.
Regarding the primes, I think they are used so that you can distinguishbetween parameters in the prior (y', m') and the current information (mperiods of observation, a total number of mE[ni] claims, and so anaggregate claim of y (please refer to eqt 9.3 in Herzog))
a final word: if i were you, i would focus more on the meaning ofparameters instead of how they look like. "meaning" can enhanceyour understanding.  ^_^

Francis


At 07:31 AM 3/27/99 , Gregor.Brown@libertynorthwest.com wrote:
>I have a question about the parameters for the Inverse-GammaDistribution.
>Can anyone humor me as to why there are primes ( ' ) on theparameters.  Is
>there any good reason or is it just done so actuarial students canuse even
>more symbol notation?
>shape parameter:  m'
>scale:  y'
>
>mean = y'/(m'-2)
>etc...
>
>Thanks,
>Gregor L. Brown
>
--=====================_2340917==_.ALT--