Sort parameters by amount of change between prior and posterior

[anders-kiaer]

Use case:

In assisted history matching (iter-0, iter-1, ..., iter-3), the user would like to see parameters that has changed the "most" when matching observations. E.g. easily find the parameters that has changed the "most" from prior (iter-0) to posterior (iter-N).

One possible solution:

In parameter matrix maybe we can offer sorting the parameters by different methods (in addition to alphabetical, we also sort by the max/min amount of "change"). There are different definitions of "changed the most". One possibility which I think could work quite well is to use entropy from information theory, see https://en.wikipedia.org/wiki/Entropy_(information_theory)

[anders-kiaer]

I think this can wait until https://github.com/equinor/fmu-dataio/issues/299 is solved.