In this repo you find useful code for selecting samples for designing listening tests.
In 'contardo' there is a state-of-art method for solving the p-dispersion problem.
In the root directory there is my code for a non mathematically-proven clustering approach to the same problem (much more efficient and much more effective). Note that my approach seems to be better here, but I don't really understand why: it seems that the Contardo's code is not suitable for this case.
To see the examples, install Gurobi, then install all dependencies for the julia files,
finally run julia -i ours.jl.
In future, there should be another approach which does p-dispersion + clustering.
Federico Simonetta https://federicosimonetta.eu.org