Open
Description
Starting point: http://recsyswiki.com/wiki/MovieLens_100k_benchmark_results
Even better if we had a free (as in software) dataset that we could ship with the software.
The idea is to record results, replay tests and compare to past runs of the suite.
This would allow us to perform quick refactoring.
We should have this for all features, including incremental updates and fold-in.