How do I create logic to merge records based on name

Help Marketplace
, ,
1

I’ve been assigned to find a solution for improving name matching in medical records. The current system processes 2-3 records at a time, identifying similar names for the bot to determine if they belong to the same person. If the similarity meets a certain threshold (currently using the Jaro-Winkler algorithm), the bot merges the records.

Challenges:

The matching process needs to handle:

  • Hyphenated last names
  • Nicknames
  • Cases where part of a name is missing

Request:

Is there a UiPath module or a more effective algorithm available that can enhance this name-matching process when Jaro-Winkler similarity score falls below a defined threshold?

2

Great topic!

I’ve used the Levenshtein distance, but according to this report the Jaro-Winkler is better.
tr-cs-06-02.pdf (anu.edu.au)
(stolen from this stack overflow post performance - Difference between Jaro-Winkler and Levenshtein distance? - Stack Overflow)

image
image350×568 40.4 KB

3

I also found this one String Similarity Algorithms (Matching Percentage) - RPA Component | UiPath Marketplace | Overview on the marketplace.

I’ll actually start using this from now on :slight_smile:

4

The report you shared was a great read! Thank you sm for your help. And yes! I have a list of modules I looked through and that is on there. Again thank you and happy friday :slight_smile:

5

This topic was automatically closed 3 days after the last reply. New replies are no longer allowed.