[Population Modeling] Discussing the Population Modeling panel in SummerSim

[Prague, Melanie]

Dear all,

To follow up with the previous emails, I will give my inputs on the three points.

1. I think that the clarification Stefan made is very important. Estimation is for me a mean to be more confident about external validity.
I would rephrase my interest in the “the importance of estimation in population modeling, compared to simulation, to ensure external validation and possibility of extrapolation/prediction of future behaviors of the system”. I have been interested a lot in epidemics (either at the cell level (for HIV) or at the population level for infectious diseases). In some cases, you want to use your “input parameters” to predict the growth of the epidemic, which would be your “output parameters” (that you may have also used for estimation). Some people use simulation to “input/set” (contrary to estimate) parameters and then predict the output. This is good regarding internal validity because if the input are correctly chosen then they will fit the observed data. But if prediction/extrapolation is of interest then there is a really need of estimation with tools such as cross-validation.

2. Separate paper is fine.

3. I will vote for short talks (10-15 min) for each panelist. Then a discussant can make a "summary talk" of 15 min trying to oppose all the ideas. Then a panel discussion for the remaining time.

Happy new year.
Best,
Mélanie.


On Jan 22, 2016, at 12:56 PM, Jeljer Hoekstra <jeljer.hoekstra at rivm.nl<mailto:jeljer.hoekstra at rivm.nl>> wrote:

Dear all,

I was invited by Jacob Barhak to join your mailinglist because I expressed an interest in the population modeling panel in Summersim.
Let me first introduce myself a little. I am a researcher at RIVM, the dutch national institute on heath and the environment. We are developing a microsimulation model of the Dutch population on the prevelence of several diseases and associated risk factors. The next few months we will be estimating parameters by calibrating the model on data of  cohort study that ran the last 25 years and for which data was measured every 5 year.
I think you can see where my interest in the panel lies.

In regard to the points Jacob raised.

1. Topic: given the small introduction you'll be not suprised that I am especially interested in estimation of model parameters as the distinction Stefan made.

2..  Short seperate paper with a common prefix would work well I think

3. I don't have a particular preference for A, B or C. But I would like to have some time to discuss what methods people have used,

best wishes
Jeljer


Proclaimer RIVM http://www.rivm.nl/Proclaimer<http://www.rivm.nl/Proclaimer>_______________________________________________
PopModWkGrpIMAG-news mailing list
PopModWkGrpIMAG-news at simtk.org<mailto:PopModWkGrpIMAG-news at simtk.org>
https://simtk.org/mailman/listinfo/popmodwkgrpimag-news

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://simtk.org/pipermail/popmodwkgrpimag-news/attachments/20160129/ae6c9e1c/attachment.html>