Phonetic Fuzzy Matching

With Phonetic Fuzzy Matching feature, you can detect and merge duplicate records with similar sounds, like "John Smith" and "Jon Smyth," ensuring your data is consistently accurate.

Follow the steps below to enable Phonetics based Fuzzy Matching to detect similar sounding duplicates:

• When configuring your desired entity, choose "Fuzzy" as your Matching Method.

• Accuracy Level: Use this field to set the phonetic fuzzy search accuracy level for retrieving duplicate records. You can choose from low, medium, or high accuracy levels:

  • Low: Detects broadly similar records, useful for catching obvious duplicates (e.g., "John Smith" and "Jon Smith").

  • Medium: Balances between precision and recall, identifying moderately similar records (e.g., "Catherine" and "Kathryn").

  • High: Finds only highly similar records, ideal for precise matching (e.g., "Michael" and "Micheal").

Note: To find duplicates with greater accuracy, we recommend selecting the high accuracy level and inputting a score above 90%.

• Score: Using this field, you can control the quality of the fuzzy search by setting the percentage value representing the fuzzy match quality. When selecting the Medium or High accuracy level, you can input the appropriate accuracy percentage in the score field.

• For medium accuracy, the range is 60-79%.

• For high accuracy, you can set the range from 80-100%.

• Once the required fields are filled in, click ‘Save and Close’, then click ‘Publish’ to publish the ‘Duplicate Matching Rule’.

• Now you will be able to detect new and upcoming duplicates using phonetic based fuzzy matching method.

• For detecting history/existing fuzzy duplicates, you need to add DeDupeD Matching Code to all the records. You can add this Code to all the existing records by using the DeDupeD Windows-based tool. This process is called Fuzzy Indexing. Kindly follow the below steps for the same:

• Login to the DeDupeD tool as System Administrator and go to “Fuzzy Indexing Tab”.

• Select the matching entity in the Fuzzy Table option for which you have configured fuzzy conditions. Click on the “Start” button to initiate fuzzy indexing.