|
|
|
The SoundsLike Phonetic Tokenizer for .Net applications is a phonetic
normalization function that converts any word to a numeric phonetic
value. It enriches your database searches and gets around the problem of
finding data that has different spellings but similar or identical
pronunciation.
SoundsLike provides two levels of phonetic conversion (narrow and wide) and
handles over 1,000 nickname and diminutive groupings (e.g. Dick = Richard)
 Database
searches will produce more matches
There are many ways to spell similar sounding names. By
incorporating the SoundsLike tokenizer your searches will become "more
intellegent" by finding more meaningful matches. Your clients will
appreciate the robust searching functionality your applications can now offer.
Simple
and easy to use
Text data in and integer values out. These numbers may be stored in your
database to querry against later. Integration is fast.
Narrow
matching
Matching is closer phonetically to the original target. Names such as Smith,
Smythe and Sammeth all return the same 32-bit integer.
Wide
matching
Matching is farther phonetically from the original target, allowing your users
to broaden their search as the fuzzy logic is more relaxed. Names such as
Smothermon and Smythers start to be included with Smith. You can give
your clients the flexibility of selecting narrow or wide search modes.
|
|
|
Nickname
groups
SoundsLike also provides a Diminutive Group for first names. Common
nicknames that belong to the same Diminutive Group will get the
same 32-bit integer. For example, Bob and Robert, Dick and
Richard, Beth / Elizabeth and Lizzie are treated as a match. SoundsLike
currently supports over 1,000 common nicknames!
Additional
scrubbing
Common typographical errors are corrected BEFORE the tokenization process
begins. Zero can sometimes be found where the letter "O" should be, like
in this example we encountered: "W00dward". SoundsLike corrects these
cases.
Several abbreviations are also exploded furthuring your matching
power. For example: 1st gets translated to First before
tokenization. This scrubbing can be disabled if desired.
Try
before you buy
We built an online demo database that includes over 43,000 first names and
170,000 last names. Each name has been tokenized by the SoundsLike Tokenizer
and the integer value outputs were stored on the same row. Try an online
demo so you can see the SoundsLike tokenizer in action!
Online Demos 
|
|
 |
|
|
|
