Creating equal lists by minimizing variance on two factors

We have a database of 773 unique words that vary based on two numerical dimensions. We would like to create 24 lists of 25 words each, with each list relatively equal on each of the two numerical dimensions. Is there a matlab function(s) that can handle this?

[Torsten]

Maybe something like

https://de.mathworks.com/help/stats/kmeans.html

?

Answer 1

[John D'Errico]

Why should there be such a function?

My guess would be you want to use a tool like kmeans, clustering your data in 2 dimensions, into 24 separate clusters. Don't expect the solution to be unique, with nice well behaved uniformly sized clusters.

If you really, desperately want 24 clusters of size 25 in each cluster, you will probably need to start writing some code of your own. But don't expect that to be easy, or result in a nice stable solution either. I'd start with k-means, and learn how to use it well, or you might want to do some serious reading about clustering techniques to write your own.