search

JeschkeLab / DeerLab-Matlab

Data analysis and method development toolbox for dipolar EPR spectroscopy
MIT License
4 stars 2 forks source link

Experiment models depend on t #84

Closed stestoll closed 4 years ago

stestoll commented 4 years ago

Currently, experiment models take the time axis t as their first input. Some of the models (such as exp_5pdeer) use this to set default parameter ranges for the refocusing times of secondary pathways, while other (such as ex_4pdeer) disregard it.

If the purpose of experiment models is to convert input parameters to the pathinfo structure, then I think they should be independent of t.

luisfabib commented 4 years ago

The ex_models must have some hard-coded starting values and bounds for the refocusing times of some dipolar pathways. Very good estimations of this starting values and boundaries can be inferred from t, especially the bounds.

I do not see right now a clean way to hard-code these boundaries without needing t, but ensuring that they still make physical sense.

The alternative would be to let all refocusing times almost unbounded (e.g. all positive-valued times), with some numerically default starting value.

stestoll commented 4 years ago

There's an analogy to the distribution models here: Good estimates for the bounds for location parameters in dd_ models would also depend on the r range. But for dd_ models, we currently do not require r to be supplied for the info = dd_mode() call.