Usage of the SoundsLike Tokenizer is very simple:
Text data goes in and integer values come out. Text data can be a first
name, last name, a medical term or any word. First names will return an
additional Diminutive Group value if the name belongs to a popular nickname
group.
Words that sound like each other will produce the same integer value. By
storing these integers on your customer record you can query against them and
return a more meaningful and intelligently matched set of records.
If the Narrow Token integers in your database match the Narrow Token value of
the search target, then the records sound like the search target.
|
|
When the Wide Token integers match, then the records loosely match
the search target.
If the Diminutive Group integer in your database matches the Diminutive Group
value of the search target, then the names are considered equal because they
belong to the same nickname group.
For example:
Bob=Robert / Abbie=Tabitha / Bill=William, etc.
* When designing logic for First Name searches, your queries should check the
Narrow Token AND Diminutive Group values so phonetic matches AND nicknames will
be included in your record set.
More 
|
|
