Data Splitting Strategies besides Random Split

Hi all,

There are numerous ways besides random split in which the interactions dataset can be split, such as temporal split, user split etc. (See https://arxiv.org/pdf/2007.13237.pdf)

So far it seems like only the random split is available as part of the lightfm.cross_validation module

Are there any plans to add these different splitting methods into the package?

Start in the lightfm.cross_validation module and review the referenced paper for temporal, user, and other interaction-splitting strategies. Clarify the intended API and semantics with maintainers; the work is done when the agreed non-random strategies are supported and their behavior is verified.

Written by the indexing model from the issue text.