Improve Cross-method-entropy to allow non-equal cluster-number instances

[Jieran-S]

Currently, there is a limitation for the cross_method_entropy script. It can only be calculated across results instances with the same number of clusters. This is due to the alignment before the entropy calculation. The results instances need to be aligned with a reference instance as the first step. To do this we use the solve_LSAP from the clue package in R, which requires both alignment parties to be of the same number of classes.

If we're able to relax this requirement by aligning the classes using other functions, we're able to relax the cross-method-entropy calculation across different methods instances, which allows for a more relaxed workflow logic. So it's a topic probably worth exploring.

Specific tasks are:

• Relaxing the align_class function for cross-method-entropy calculation
• Relax the input file format/logic to allow for flexible script input