yannikschaelte / hieropt Goto Github PK

Implement standard options. Standard for everything should be to have all grouped together and with absolute values (then the users have to specify only stuff that they want to do hierarchically).

While Jakstat works, Rafmekerk does not return the desired results. This might be due to some wrong handling of experimental conditions.

There should be at least 3 test functions, e.g. jakstat_swameye, rafmekerk, erbb. For all of these, implement the optimization pipeline and perform and store (.mat) test results to enable a foundated comparison of the different approaches.

offset parameters might be problematic here, but proportionality factors will work

The computation of the optimal parameters depends on a few assumptions. These should be checked, as well as validity (dimensions etc.) of the input options and data.

currently, only 0/1 default values are accepted. The approach however also works with any user-defined values, which it should be possible to pass via a ne1 cell array of ntny*nr matrices, for b, c and noise.