Function vtext::metrics::string::dice_similarity

pub fn dice_similarity(x: &str, y: &str) -> f64

Sørensen–Dice similarity coefficient

This similarity tokenizes the input string x, y as 2-char n-grams, into two sets of tokens X, Y then computes,

similarity(x, y) = 2 * |X ∩ Y| / (|X| + |Y|)

where |X| is the cardinality of set X.

Example

use vtext::metrics::string::dice_similarity;
use float_cmp::ApproxEqUlps;

let res = dice_similarity("yesterday", "today");
assert!(res.approx_eq_ulps(&(1./3.), 4));