Does anyone know how reclink chooses which potential matches to report? Does it effectively sort the potential matches by similarity score and start with the match with the highest score (greedy algorithm)? Or does it use some sort of optimal algorithm? Or something else?

Similarly, for people who use matchit, how do you choose which potential matches to use when doing a 1:1 fuzzy match of two datasets?

I'm looking more for best practices than code, though I'd be interested in code that maximized the total similarity score if anyone had such a thing.

Thank you,

Kramer