PDA

View Full Version : 'Duplicate' Search Results


siliconkibou
01-13-2004, 01:08 AM
Hi,

I've noticed that PHPDig seems to not be able to differeniate between nearly identical(I say nearly, because they appear identical to my human eyes) documents located on a website.

If one document is located in say /worldwide/ and another in /about_us/ they both come up in a search result with identical percentages.

Additionally, documents that are generated dynamically but are identical also give multiple duplicate results.

For example:

http://www.issa.com/worldwide/index.jsp?region=9&type=news&id=153

and

http://www.issa.com/worldwide/index.jsp?region=11&type=news&id=153

Both are listed as results(they differ by the region variable in the URL).

This behavior is understandable, since they are slightly different(from a machines perspective).

However, is there a way to increase the criteria used to judge duplicate documents to filter out highly similar documents as well?

Say if they share 90% of the same content?

Thanks in advance,

-Paul

For reference, you may see for yourself this behavior at:

http://search.custodialadvisorsnetwork.org

Search for "cleaning standards" as a good example.

Several pages into the search, you'll see some examples of pseudo-duplicates.

Charter
01-13-2004, 09:00 AM
Hi. You might try modifying the $md5 variable talked about in this (http://www.phpdig.net/showthread.php?threadid=242) thread.