Rosette can compare two entity names (Person, Location, or Organization) and return a match score from 0 to 1.
Rosette handles name variations (such as misspellings and nicknames) with multilingual support by using a linguistic, statistically-based system.
When you submit a name-similarity request, you must specify name1 and name2 with the text for each. Although not required, we strongly recommend specifying the source language (if known) for better accuracy. If you do not specify the source language, then the Rosette name similarity algorithm will guess the language.
Tip: If you know the language of your names, supply that value as the
language parameter for each name object. That will improve your results' accuracy.
Note: Do not use the Rosette language detection endpoint to determine source language of names as the language guesser within the name similarity endpoint uses an algorithm tuned specifically for names rather than general text.
If you'd like to try out Rosette's pairwise matching name capabilities, sign up for a free trial of Rosette and use the
The Name Similarity REST endpoint is: