Senzing strives to handle the data you present to it and has always done an excellent job of automatically identifying and correcting for overused values that become unreliable to help identify entities. Much of this is domain aware handling of invalid, misused or abused data attribute values.
Senzing has always been the industry leader in handling multi-cultural names in entity resolution, we continue extending upon that success with further enhancements to our comprehension of entity based attributes.
US Social Security Numbers
It's highly likely you recognize invalid SSNs such as 111-11-1111 and 123-45-6789. But do you recognize 219-09-9999 and 078-05-1120 were valid SSNs used between 1938-1940 in advertisements and became two of the most abused SSNs in history?
Prior to 2011 and the US Social Security Number Randomization project you could even reasonably validate that an SSN had been issued using the periodically distributed High Group List. After 2011, the SSNs become harder to validate but Senzing is able to leverage the restrictions still in place to improve the automatic domain aware handling of SSNs.
What happens if an SSN is found to be invalid, misused or abused?
- First, Senzing will validate the SSN as a match, but not with the same level of importance as a valid SSN. Additional information will be required to make a match.
- More importantly, Senzing will not keep records apart simply because of differing SSNs where one or both is invalid. This means when you refused to give your SSN and the computer system or operator entered 999-99-9999 as a place holder, your records won't be penalized.
IDs in General
IDs are an area Senzing will continually work to improve, as an almost infinite domain worldwide. Today Senzing recognizes the most common issues with generic IDs, for example:
- Comprised of one repeating digit (e.g. 1111111)
- It is valid to have multiple IDs of the same type in multiple domains without a conflict (e.g. driver's licenses issues in California and Nevada or passports from the USA and Germany)