Consider partition data reduction algorithm

[malcolmbarrett]

the Partition algorithm has several nice properties, particularly in high dimensions. It's also completely reproducible without a seed. https://github.com/USCbiostats/partition

[EmilHvitfeldt]

Correct me if I'm wrong, but isn't this method closer to a dimensionality reduction method?

[malcolmbarrett]

Now that you mention it, the clustering is really happening at the column level, not the row level like this package is focused on. Maybe better suited to a recipe somewhere?

[EmilHvitfeldt]

If a trained application can be reapplied to new data, then it would be perfect as a recipe step!