ActEd  

Go Back   ActEd > Subject Discussions > CT Subjects > CT4
Reply
 
Thread Tools Display Modes
Old
  (#1)
payal modi payal modi is offline
Newcomer
 
Posts: 1
Join Date: Mar 2009
Default Exposed to risk - 12-03-2009, 05:55 PM

Can anyone tell me the basic difference between initial exposed to risk and central exposed to risk? I mean to ask if we say that tha IETR for a person is 11months and CETR is 7months then what actually mean by this two numbers.........Please do reply as early as possible.

Thanks in advance!!

Last edited by payal modi : 14-03-2009 at 04:47 AM.
Reply With Quote
Old
  (#2)
jensen jensen is offline
Forum FIFA
 
jensen's Avatar
 
Posts: 519
Join Date: Jan 2008
Location: 3rd rock from sun
Default 15-03-2009, 05:48 AM

They're just two different ways of measuring the length of time a person is exposed to the risk of dying.

If Initial E = 11 months, then Central E should be 11-6months = 5.

What I would like to know is, why do we use Initial E in the binomial model and Central E in Poisson model?
Reply With Quote
Old
  (#3)
MarkC MarkC is offline
Frequent user
 
Posts: 14
Join Date: Apr 2008
Default 15-03-2009, 01:33 PM

ActEd's "lightbulb boxes" in Chapter 10 prove quite helpful here...

Suppose you observe a life over the age interval (x+a, x+b), with 0 < a < b < 1.

Suppose the life survives to age x+b. You've observed the life for (b-a) years. It contributes (b-a) to both the initial and central exposed to risk.

Now suppose the life dies at age x+t, with a < t < b. You've observed the life for (t-a) years.
  • It contributes (t-a) to the central exposed to risk...
  • but it contributes (1-a) to the initial exposed to risk.

So the central exposed to risk is the actual time you spend observing lives. The initial exposed to risk has some sort of "adjustment" in respect of those lives observed to die.
Reply With Quote
Old
  (#4)
didster didster is offline
Forum FIFA
 
Posts: 523
Join Date: Apr 2007
Default 15-03-2009, 11:32 PM

If I remember correctly, the reason we use two different Exposed to Risk is because each is suited to a different model. The exposed to risk is determined by what is appropriate for a given model.

Poisson model uses the central exposed to risk.
It is a continuous distribution with the "probabilities" (read force of mortality) apply at each INSTANT to the exposed lives at that instant.
Decrease the number of lives (say if some die) then the exposed lives reduces. If you think back to the derivation of formulae for the poisson process, you always consider loads of instants in time and add them up.
So we need the "exact time" to calculate the exposure.

You can think of the exposure being an average for the period, which may be a bit like the exposure at the middle of the period, hence the term central exposed to risk.

Binomial model looks at number of exposed lives at the START and counts how many are alive at end of period. It is discrete, either you die or not in the period, and doesn't matter if it's near the start or the end of the period.
So an INITIAL exposed to risk is used.
Reply With Quote
Old
  (#5)
jensen jensen is offline
Forum FIFA
 
jensen's Avatar
 
Posts: 519
Join Date: Jan 2008
Location: 3rd rock from sun
Default 19-03-2009, 10:01 AM

Quote:
Originally Posted by didster View Post
If I remember correctly, the reason we use two different Exposed to Risk is because each is suited to a different model. The exposed to risk is determined by what is appropriate for a given model.

Poisson model uses the central exposed to risk.
It is a continuous distribution with the "probabilities" (read force of mortality) apply at each INSTANT to the exposed lives at that instant.
Decrease the number of lives (say if some die) then the exposed lives reduces. If you think back to the derivation of formulae for the poisson process, you always consider loads of instants in time and add them up.
So we need the "exact time" to calculate the exposure.
.
Thanks didster

I understand the binomial one, but not Poisson. I thought poisson is a discrete process? If it can be approximately Normal, then the same can be said for the Binomial model.
Reply With Quote
Old
  (#6)
didster didster is offline
Forum FIFA
 
Posts: 523
Join Date: Apr 2007
Default 19-03-2009, 11:21 AM

I'll try to be a bit more specific, although it's difficult since it's been a while.

The survival model is the Poisson process, which is not to be confused with the Poisson distribution.

Poission process is a discrete state, continuous time model.
Here we could have the states being the number of lives present.

The Poission distribution is the distribution of a random variable, being the number of states that you move. The number of states is discrete, and Poisson distribution is discrete random variable.

So the movement is continuous (in time), but the counting of the movements is discrete. That's the link between the two "Poisson"

The other link with the Poisson process is that the times between jumps is an exponential distribution (and each time is independant).

I'm sure there is something in the notes about the poisson process. Have a read of that (including the maths involved), then try to apply such a model to lives, thinking about how to measure the Exposed to Risk.
Reply With Quote
Old
  (#7)
jensen jensen is offline
Forum FIFA
 
jensen's Avatar
 
Posts: 519
Join Date: Jan 2008
Location: 3rd rock from sun
Default 19-03-2009, 12:50 PM

Thanks didster.

I did read about Poisson process, but I failed to see the link between that and the whole survival model. Maybe that's why I didn't get it.

Will read it again and try to see the big picture.
Reply With Quote
Reply


Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

vB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump



Powered by vBulletin® Version 3.6.4
Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
vBulletin Skin developed by: vBStyles.com