Closed AuroreAA closed 1 year ago
I would maybe prefer the functions ICSclust and PCAclust. And in general it would be nice to
However looking at this it would require many many parameters to be passed on, or? Would it be then easier to have one general function with few parameters but does all more or less automatic and then we have the separate building block functions in case the user wants to "fine tune"?
I generally agree that this will be nice to have when we release the package. It's not urgent for doing the simulations for our paper, where we need to do the steps separately to avoid repeated computations.
@klauschN, can you clarify what you mean by point 4? That is not clear to me.
Indeed, we certainly also need separate functions for separate steps at the user level, but this is what @AuroreAA has implemented now. I think having a broader function that does everything in one go will be nice, but there are some challenges.
Ah typing error in 4. I thought that either the method gives the clustering for a specific number of clusters or is able to decide itself on the number of clusters.
Do we want to have more general functions, for example
tandemClust
orICSclust
? if yes, what do we want to provide as functionalities to the user with which amount of parameters?