fast-dice-coefficient
Fastest implementation of Sørensen–Dice coefficient. This implementation has linear time complexity O(n), as opposed to other solutions: string-similarity, dice-coefficient, etc., running in quadratic time O(n2).
This algorithm can be used to compute the similarity between strings. It returns a fraction between 0 and 1, where 0 indicates no match and 1 indicates full match.
Install
$ npm install fast-dice-coefficient --save
Usage
var dice = require('fast-dice-coefficient');
console.log(dice('javascript', 'coffeescript'));
//=> 0.5
Benchmark
200 samples on i7-3667u with 8GB RAM.
Using two test-strings with 100 characters each.
First 4 packages uses dice-coefficient and last uses levenshtein.
Package | ops/sec |
---|---|
stringSimilarity | 7,590 |
natural.DiceCoefficient | 11,117 |
dice-coefficient | 14,732 |
fast-dice-coefficient | 44,067 |
fast-levenshtein | 4,495 |