The purpose of this article is to explain the possibility for a Member Association to customize person deduplication process by setting a score threshold.
This feature should be considered by Member Associations having a high number of low scored potential duplicates. Such organisations might struggle with a longer registration/approval process as operators need to review more records.
Each match (potential duplicate) returned by FIFA Connect ID has its calculated score. Score represents how accurate a match is comparing to the original (input) data. The simplest interpretation of the score is that the higher it gets the more accurate data could be expected. Score equal to 1 means that an exact match has been found.
Example 1 (no threshold)
An attempt to register John Smith resulted in three matched records:
The first match turned out to be exact match with score 1. Other potential duplicates were found with scores 0.95 and 0.65. Operator of the system has to consider all three records when resolving the registration.
FIFA Connect ID offers each Member Association a possibility to customize duplicate resolution algorithm. Some of the matches could be automatically filtered out by setting a custom score threshold. This will result in all matches below the threshold being ignored during the deduplication process. Only matches with scores equal or above the threshold will be taken into account (and returned) while attempting to register a player.
Default value of the threshold is 0 which means that no matches are being filtered out. Threshold set to 0.7 means that only matches with scores equal or higher than 0.7 will be returned as potential duplicates.
Example 2 (with a threshold)
Let's consider the example above once again but with a threshold set to 0.7. Another attempt to register the player will result in the following potential duplicates found:
Duplicate with score 0.65 has been omitted as it is below our custom threshold.
Note that the higher the threshold:
- the fewer potential duplicates will be returned
- the less frequently will you need to use ForceRegister(...)
- the faster will your registration process be
On the other hand with a higher threshold a risk of registering a duplicate is greater, which decreases the data quality in FIFA Connect ID Service. For that reason we do not recommend using a threshold equal or higher than 0.6. Also, we recommend to let the players registration process run for some time and only then, having feedback from operators, decide about the custom value.
To set the custom threshold for your organisation contact FIFA Connect ID support team.