You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In some cases it can be interesting/informative to compare the fits to the real data to those from simulated data. I.e., how easy it is to identify the real data from 9 other simulated data sets (Buja et al. 2009; https://doi.org/10.1098/rsta.2009.0120).
I propose a new argument to run_osa called simulate which will replace the observed data with simulated ones, given the model expectations.
Then a wrapper can be developed called run_osa_simulator(..., nsims=n) or something like that, which will return n+1 fits for plotting, one being the real data and the other n the simulated data. The user can then quickly visualize how different the real data looks from simulated ones matching the model.
Thoughts?
The text was updated successfully, but these errors were encountered:
In some cases it can be interesting/informative to compare the fits to the real data to those from simulated data. I.e., how easy it is to identify the real data from 9 other simulated data sets (Buja et al. 2009; https://doi.org/10.1098/rsta.2009.0120).
I propose a new argument to
run_osa
calledsimulate
which will replace the observed data with simulated ones, given the model expectations.Then a wrapper can be developed called
run_osa_simulator(..., nsims=n)
or something like that, which will return n+1 fits for plotting, one being the real data and the other n the simulated data. The user can then quickly visualize how different the real data looks from simulated ones matching the model.Thoughts?
The text was updated successfully, but these errors were encountered: