All candidate pairs with their per-field comparison vectors

Like candidate_scores(), but returns the raw per-field similarity columns (the graded comparison vector) rather than a single combined score. This is the substrate for fitting a match model (fit_match_model()).

Usage

candidate_comparisons(df, cols = synthid_cols(), features = NULL)

Arguments

df

A raw panel (as for link_panel()).

cols

Column mapping; see synthid_cols().

features

Feature columns to compare (defaults to the standard set).

Value

A data frame, one row per candidate pair: row_x, row_y, yr_x, yr_y, org, surname_w_x, surname_w_y, and one numeric similarity column per feature (NA where a field was missing on either side).