MarkLogic - Person name fuzzy match

Question

We are using MarkLogic version 9.0-6.2

We have a requirement to match person names based on some fuzzy logic (For example, Mike and Michael are same). Each person has a first and last name. What is the best way to achieve this in MarkLogic. Does smart mastering help?

We do not have to merge the matching documents. If a match is found, we need to return a message.

Answer 1

Yes, you can use Smart Mastering for this. I'd use thesaurus for first name and double metaphone for last name. Smart Mastering does allow you to call just the match functions and skip merging.

Answer 2

I would look into using spell.doubleMetaphone for this, that gives pretty strong fuzziness already. It won't cause Mike (mk) and Michael (mkl, mxl) to match, but you can potentially combine it with thsr.lookup on a custom defined thesaurus with name synonyms. That together should give you a pretty powerful fuzziness.

HTH!