Prev: sas time value
Next: Schemper's V macro sas 9.2
From: Chris T on 30 Oct 2009 16:55
Maybe my subject isn't great, but here's the situation: I have lots of
strings, some of them may be similar to another. Is there something
out there that (magically!) calculates the similarity, or maybe a
match percentage between two strings?

I have some theories on how to do this programatically, but would
rather use something that's already been tested.
From: Arthur Tabachneck on 30 Oct 2009 18:03
Chris,

Take a look at the functions shown in the following thread:
http://groups.google.com/group/comp.soft-sys.sas/browse_thread/thread/bd902a1d31fb7a84

HTH,
Art
--------------
On Oct 30, 4:55 pm, Chris T <chris.tin...(a)gmail.com> wrote:
> Maybe my subject isn't great, but here's the situation: I have lots of
> strings, some of them may be similar to another. Is there something
> out there that (magically!) calculates the similarity, or maybe a
> match percentage between two strings?
>
> I have some theories on how to do this programatically, but would
> rather use something that's already been tested.
From: Patrick on 31 Oct 2009 00:10
Besides this interesting thread Art refers to:

There is also the "SAS Data Quality Server " out there which alows you
to calculate match codes (i.e. on 95% probability)
http://support.sas.com/documentation/whatsnew/91x/dqclrefwhatsnew900.htm.

Sure worth to have a look into it in case your site has got a license.

HTH
Patrick
From: Sigurd Hermansen on 3 Nov 2009 15:04
On http://sconsig.com/ (tip 00342) shows an extended example of the use of the SAS SPEDIS() function (the function that resolves certain spelling errors in SAS programs) when comparing strings for similarity score. We use this "cost of rearranging string1 to match string2" function for many purposes.
S

-----Original Message-----
From: SAS(r) Discussion [mailto:SAS-L(a)LISTSERV.UGA.EDU] On Behalf Of Chris T
Sent: Friday, October 30, 2009 4:56 PM
To: SAS-L(a)LISTSERV.UGA.EDU
Subject: Comparing strings - assigning a similarity rating

Maybe my subject isn't great, but here's the situation: I have lots of
strings, some of them may be similar to another. Is there something
out there that (magically!) calculates the similarity, or maybe a
match percentage between two strings?

I have some theories on how to do this programatically, but would
rather use something that's already been tested.
 | 
Pages: 1
Prev: sas time value
Next: Schemper's V macro sas 9.2