search

andrewhooker / PopED

Population Experimental Design (PopED) in R
https://andrewhooker.github.io/PopED/
GNU Lesser General Public License v3.0
33 stars 21 forks source link

Manually updating the number of samples as an alternative to opt_samps in R? #42

Closed michielve closed 4 years ago

michielve commented 4 years ago

Hello,

Thank you for your work on PopED. As with issue #40, I am trying to optimize the number of samples using a function that is not yet available in the R version of PopED.

Before this function gets implemented, I am wondering whether it would be a correct approach to run multiple optimization procedures while testing different designs (e.g. with 4 to 20 samples per subject) and judging the output?

If this is possible, what is the best approach to compare the different designs since more samples will always provide more information? Can this comparison be done by comparing the OFV? Plotting the mean RSE of the parameters over the number of samples? Determining a 'significant' change in the efficiency?

Thank you for your suggestions,

Michiel van Esdonk

andrewhooker commented 4 years ago

Hi Michiel,

I think your strategy makes sense. Adding samples will always give more information and you will then need to determine when you are satisfied with the amount of information. As you suggest, using a cutoff level for OFV, efficiency or RSE of certain parameters may be appropriate, and will be case specific I believe.

Andrew