Name-key frequency table for a reference corpus

Counts how many reference records share each exact name_key. A name that occurs once or twice is distinctive (an exact match to it is trustworthy even without address corroboration); a name shared by many records (e.g. "FIRST BAPTIST CHURCH") is not. Used by the distinctive-exact-name promotion in np_score(method = "hier"). This is more robust than summed token IDF, which conflates distinctiveness with name length.

Usage

np_name_freq(name_keys)

Arguments

name_keys

Character vector of normalized names (reference name_key).

Value

A named integer vector mapping name_key -> record count.