[AccessD] "fuzzy logic" search

John Colby jcolby at colbyconsulting.com
Tue Oct 21 20:05:44 CDT 2003


In the end, they want to be able to type garbage in and have the database
figure out what they mean.  I assume you have never actually talked to
users?  ;-)

John W. Colby
www.colbyconsulting.com

-----Original Message-----
From: accessd-bounces at databaseadvisors.com
[mailto:accessd-bounces at databaseadvisors.com]On Behalf Of David McAfee
Sent: Tuesday, October 21, 2003 8:33 PM
To: 'Access Developers discussion and problem solving'
Subject: RE: [AccessD] "fuzzy logic" search


I was going to recommend soundex on the names as well. Soundex is pretty
cool, but also makes some matches that are not that close.

It sounds like someone is using natural keys where they shouldn't be? :P

(that might start a war)

Maybe they can search for duplicate claims for a given address, name or type
of injury and name?

David

-----Original Message-----
From: accessd-bounces at databaseadvisors.com
[mailto:accessd-bounces at databaseadvisors.com]On Behalf Of Charlotte
Foust
Sent: Tuesday, October 21, 2003 5:20 PM
To: Access Developers discussion and problem solving
Subject: RE: [AccessD] "fuzzy logic" search


John,

I've received this kind of request before and I looked into it once.
The fuzzy logic math was way, way beyond my grasp, so I settled for
offering alternatives like soundex on the names.  I created a customized
soundex that used a number for leading vowels rather than the standard
"vowel followed by number" pattern, so that if they misspelled a name
with a leading vowel, there were still possible matches.  But SSN?  I
mean, think of it.  If a SSN has 9 digits, any one of which can repeat,
how do you determine a mismatch by up to 2 characters?  WHICH 2
characters?

Charlotte Foust

-----Original Message-----
From: John Colby [mailto:jcolby at colbyconsulting.com]
Sent: Tuesday, October 21, 2003 4:06 PM
To: AccessD
Subject: [AccessD] "fuzzy logic" search


My client is having difficulties where the data input people are
entering duplicate claims because wrong names or SSNs are entered.  He's
asked how difficult it would be to do a search for the names entered
(seems irrelevant if it is wrong) but then has asked how difficult it
would be to do a "fuzzy logic" search for SSNs that are different by up
to 2 characters.  I don't even have a clue how to attack something like
that.  Any ideas (other than "out the door, no parachute")?

John W. Colby
www.colbyconsulting.com


_______________________________________________
AccessD mailing list
AccessD at databaseadvisors.com
http://databaseadvisors.com/mailman/listinfo/accessd
Website: http://www.databaseadvisors.com
_______________________________________________
AccessD mailing list
AccessD at databaseadvisors.com
http://databaseadvisors.com/mailman/listinfo/accessd
Website: http://www.databaseadvisors.com




More information about the AccessD mailing list