Strings enable you to create “rules” or “guidelines” that instruct the algorithm on how to handle certain incoming data values.
For example, if entering member data when not all of the member information is available, “dummy” data may be entered in some required fields (such as 999999999 for a Social Security number or 5555555555 for a phone number). If this same dummy data were entered for every member in which the Social Security number was not available, you could end up with multiple records having the same 999999999 number and thus, causing inaccurate comparison scores. By defining these values in an anonymous string, the algorithm knows to ignore these values during bucketing and comparison.
Another example of string use is in mapping specialty codes or professional titles that can be part of a person or business name. Standardization routines are used to standardize member names (first, last, middle); however these names can often include additional words or titles that can assist in achieving more accurate comparison and identification. Abstracts are a means of providing this additional information when comparing and identifying members. To use the abstracts in bucketing and comparisons, the values must be standardized. The result of mapping is the “abstract value.” Abstracts are defined codes that map to specific words found in member data which represent the same thing. The words “Heart” and “Cardio,” while not the same, can represent the same specialty. Abstracts might be inferred from two places: 1) from words embedded in the member name, or 2) from specific attribute values. For example, a comparison of the following records:
“J Smith” with a specialty code attribute of “Cardiologist”
“J Smith Heart Center”
“J Smith” with a specialty code attribute of “Dentist”
“Cardiologist” and “Heart” both map to an abstract value of “CRD,” whereas “Dentist” maps to an abstract of “DNT.” The abstract comparison between “J Smith, Cardiologist” and “J Smith Heart Center” adds a small amount to the score between these members to distinguish it from comparisons with “J Smith, Dentist.”