Showing results for 
Search instead for 
Did you mean: 

Scaling factor for apoc.text.jaroWinklerDistance

According to the formula for Jaro-Winkler Distance (Jaro–Winkler distance - Wikipedia), there is an adjustable scaling factor that can give more favourable ratings to strings that matches from the beginning.

However, in Neo4j, the only input parameters for apoc.text.jaroWinklerDistance() are just the two strings we would like to compare (apoc.text.jaroWinklerDistance - APOC Documentation)

  • What is the standard scaling factor used for APOC's Jaro-Winkler function?
  • Is there a way to adjust this scaling factor in APOC?

Graph Maven

Per Winkler's original article, the standard value is 0.1. I did some manual calculations using the formulas set for Jaro-Winkler similarity and the results compared well with Neo4j algorithm.