Hi,
I try to calibrate parameters in my model and have successfully calibrated the variables using Optimization Setup based on one set of time-series data.
However, I need to use 51 sets of data and make the calibration more reasonable. Can I use the subscripts, Trial: (T1-T51), to import data and then calibrate parameters using Optimization Setup? Is there any other methods to perform calibration?
Thanks for your help.
Emma
Question about Model Calibration
Re: Question about Model Calibration
There are generally two strategies.
- You could create 51 datasets for your model and estimate each individually, or
- You can subscript the model and estimate everything at once. (I think this is what you are suggesting.)
I think the second approach is generally preferable, unless you have computation speed problems.
You can add the [trial] dimension to parameters, but it generally makes sense to pool them as much as possible, to increase the sample size on which each is based, and minimize the size of the search space.
- You could create 51 datasets for your model and estimate each individually, or
- You can subscript the model and estimate everything at once. (I think this is what you are suggesting.)
I think the second approach is generally preferable, unless you have computation speed problems.
You can add the [trial] dimension to parameters, but it generally makes sense to pool them as much as possible, to increase the sample size on which each is based, and minimize the size of the search space.
/*
Advice to posters (it really helps us to help you)
http://www.ventanasystems.co.uk/forum/v ... f=2&t=4391
Blog: http://blog.metasd.com
Model library: http://models.metasd.com
Bookmarks: http://delicious.com/tomfid/SystemDynamics
*/
Advice to posters (it really helps us to help you)
http://www.ventanasystems.co.uk/forum/v ... f=2&t=4391
Blog: http://blog.metasd.com
Model library: http://models.metasd.com
Bookmarks: http://delicious.com/tomfid/SystemDynamics
*/