Hi experts,
I’d like to try to estimate the parameters of the commodity model included in the book Business Dynamics by Prof. John Sterma (the general one applied to pulp and paper industry) with copper data.
I’ve collected some data from various government sites in USA, even though some data are difficult to find.
What’s the best approach for calibrating the parameters? Can I use the optimization engine included in Vensim in order to fit the parameters to real data?
I’m not sure tha Kalman filter is the appropriate technique.
Ore another strategy is to estimate separately the parameters, using regression, etc. ?
Any advice will be appreciated, especially from someone who knows already that model.
Many thanks in Advance
G
Calibration of parameters in SD models
-
bob@vensim.com
- Senior Member
- Posts: 1107
- Joined: Wed Mar 12, 2003 2:46 pm
Calibration is very tricky with cyclic behavior and if you want to do a full estimation of parameters you will need to use Kalman filtering or a related technique to get state estimates. A good reference is: "Statistical Tools for System Dynamics" by D.W. Peterson in Elements of the System Dynamics Method edited by Jørgen Randers (Chapter 11, pp. 224-241, The MIT Press, Cambridge MA, 1980).
If you have sufficient data, both in terms of the number of variables measured and the consistent reporting of values at times, you can run a regression. Equivalent to this you can adjust the model to use the data as exogenous inputs, in this case filtering is not necessary.
If you have sufficient data, both in terms of the number of variables measured and the consistent reporting of values at times, you can run a regression. Equivalent to this you can adjust the model to use the data as exogenous inputs, in this case filtering is not necessary.