Jaro–Winkler Similarity Calculator
Calculate Jaro-Winkler similarity by boosting Jaro similarity with a common-prefix bonus of scale 0.1 per character up to four characters, applied when Jaro similarity exceeds 0.7.
Description
Calculate Jaro-Winkler string similarity with a prefix bonus for closely matching short strings.
Jaro-Winkler similarity extends Jaro similarity by giving closely matching strings a small boost when their first characters agree. The measure is often used for personal names and other short identifiers where the beginning of the string is especially informative.
When to use Jaro-Winkler Similarity Calculator
- Rank likely matches among short names
- Compare the effect of a common prefix on Jaro scores
- Support record-linkage experiments with an interpretable string metric
How the calculation works
The calculator first obtains Jaro similarity. When that score exceeds 0.7, it adds 0.1 × prefix length × (1 − Jaro), with the common prefix capped at four characters. Below the threshold, the result remains the original Jaro score.
Interpreting the result
The output shows both the boosted score and its unboosted Jaro basis. `commonPrefixLength` explains the maximum possible bonus, although no bonus is applied when Jaro similarity is 0.7 or lower. Compare scores only under the same preprocessing rules.
Important limitations
- Prefix emphasis may be inappropriate when suffixes or middle tokens carry more identifying value.
- Comparison is case-sensitive and does not normalize Unicode or whitespace.
- A high score is not proof that two records refer to the same entity.