|
|
|
|
|
by cjk2
713 days ago
|
|
Not research, but I spent 5 years writing matching algorithms for data for financial companies which was entered from random crappy mainframes driven by the lowest bidding data entry workers. This tied together the financial relationships of millions of people. The dumber it was the measurably more accurate it was. Hell even some of it used soundex scoring and a naive z-test to work out the confidence that two names were matching. |
|