Open p-lauer opened 5 years ago
In GitLab by @hehnen1 on Nov 29, 2017, 11:42
This raises furthermore the question, how to evaluate the fitness in general. One could imagine two sets of simulation results yield a very similar fitness value. However, when plotted in comparison with experimental data one might "visually" be closer then the other. This would depend on the evaluation method chosen, of course.
Should this be a new issue, since it is a very fundamental question?
In GitLab by @hehnen1 on Dec 17, 2017, 19:28
added ~276 label
In GitLab by @hehnen1 on Nov 29, 2017, 11:37
Use a range around the best fitness value, to extract a group of parameter sets the fitness values of which are located within sayed range. Idea is to assess the variation of the simulation results for the parameter sets that are close to the "best" set.
Connected to issue #27