fuzz_functions {revtools}

Functions for fuzzy string matching

Description

Duplicate of functions from the Python library 'fuzzywuzzy' (https://github.com/seatgeek/fuzzywuzzy). These functions have been recoded from scratch based on the description given here. For consistency with stringdist, however, these functions are computed as distances rather than similarities; i.e. low values signify similar strings.

Usage

fuzzdist(a, b, method)
fuzz_m_ratio(a, b)
fuzz_partial_ratio(a, b)
fuzz_token_sort_ratio(a, b)
fuzz_token_set_ratio(a, b)

Arguments

a

a string

b

a vector containing one or more strings

method

a function to be called by 'fuzzdist'

Value

A score of same length as y, giving the proportional dissimilarity between x and y.

Note

fuzz_m_ratio is a measure of the number of letters that match between two strings. It is calculated as one minus two times the number of matched characters, divided by the number of characters in both strings.

fuzz_partial_ratio calculates the extent to which one string is a subset of the other. If one string is a perfect subset, then this will be zero.

fuzz_token_sort_ratio sorts the words in both strings into alphabetical order, and checks their similarity using fuzz_m_ratio.

fuzz_token_set_ratio is similar to fuzz_token_sort_ratio, but compares both sorted strings to each other, and to a third group made of words common to both strings. It then returns the maximum value of fuzz_m_ratio from these comparisons.

fuzzdist is a wrapper function, for compatability with stringdist.

Examples

fuzz_m_ratio("NEW YORK METS", "NEW YORK MEATS")
fuzz_partial_ratio(
  "YANKEES",
  c("NEW YORK YANKEES", "something else", "YNAKEES")
)
fuzz_token_sort_ratio("New York Mets vs Atlanta Braves", "Atlanta Braves vs New York Melts")
fuzz_token_set_ratio(
  a = "mariners vs angels other words",
  b = c("los angeles angels of anaheim at seattle mariners", "angeles angels of anaheim ")
)