Skip to contents

Fuzzy Token Set Ratio

Source: R/text.R
fuzzy_token_set_ratio.Rd

This function computes a fuzzy similarity score between two strings based on the token set ratio methodology. It considers the intersection and differences between tokenized word sets from the input strings, and calculates a similarity score normalized by string lengths.

Usage

fuzzy_token_set_ratio(s1, s2, score_cutoff = 0)

Arguments

s1

A character string. The first string to compare.

s2

A character string. The second string to compare.

score_cutoff

A numeric value (default is 0) specifying the minimum similarity score threshold. Scores below this threshold may trigger early exits in the computation.

Value

A numeric similarity score between 0 and 100, representing the degree of similarity between the two input strings.

Details

This function performs the following steps:

  • Tokenizes the input strings.

  • Identifies intersecting and differing tokens between the two tokenized sets.

  • Computes the longest common subsequence (LCS) distance for differing tokens and normalizes it.

  • Calculates similarity ratios for intersecting tokens combined with differing token sets.

  • Returns the maximum of the normalized LCS distance and the two intersecting token ratios.

The function short-circuits to return 100 if one token set is a subset of the other. If either input string is empty, the function returns 0.

Examples

# Example usage:
fuzzy_token_set_ratio("fuzzy was a bear", "fuzzy was a dog", score_cutoff = 80)
#> [1] 84.61538
fuzzy_token_set_ratio("hello world", "world hello")
#> [1] 100