Kingsley,
The issue is not so much “noise” – although there is certainly some of that (erroneous and duplicative data)- as it is the multiplicity of contexts/perspectives implicit in the data and the varying/overlapping scope of said contexts and associated perspectives that make this a difficult problem. Disambiguation and “de-duplication” is accomplished to varying degrees by teasing out the implicit context/perspective and scope assumptions in the data through application of test/trial markers (your “relation/facet filtering criteria”) for common and plausible context and scope assumptions. Needless to say, there is no guarantee that one will get this right for any given arbitrary data collection. Nor can we know what the data collection does not include (that might be relevant to our purposes for accessing the data collection in the first place), unless, of course, the contexts and scopes of the collection are made explicit.
Hans
From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx [mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx] On Behalf Of Kingsley Idehen
Sent: Tuesday, June 18, 2013 11:46 AM
To: ontolog-forum@xxxxxxxxxxxxxxxx
Subject: Re: [ontolog-forum] How de facto standards are created
On 6/18/13 10:55 AM, David Eddy wrote:
Kingsley -
On Jun 17, 2013, at 10:41 AM, Kingsley Idehen wrote:
[1] http://lod-cloud.net/ -- lod cloud pictorial circa. 2011 (we are now way over 50 Billion triples and on an exponential growth curve)
Is there any practical value in 50 billion triples?
Of course.
It's been my experience that the more stuff one combines from more sources, the noise level just goes through the roof.
So you need entity disambiguation as a feature of tools that interface with the LOD cloud.
Or have all the data in all the sources in that eye-chart been vetted? Meaning: How do I know that "MA" in my source means the same as "MA" in your source?
Good question, some answers:
1. http://bit.ly/11Mjx5s -- placement in a drill-down from which you can disambiguate "MA" based on relation/facet filtering criteria.
2. http://bit.ly/15OQmmf -- 'New York' disambiguated
3. http://bit.ly/13IM8u9 -- 'Paris' disambiguated
4. http://bit.ly/12gFJqK -- 'Parker' disambiguated .
You can never have too much data, you just need the right tools for drill-down based insight discovery. In addition, as we move to a more sensory Internet, expect to unintentionally sneeze a billion+ triples as part of your combined Web and Internet experience :-)
Kingsley
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
--
Regards,
Kingsley Idehen
Founder & CEO
OpenLink Software
Company Web: http://www.openlinksw.com
Personal Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter/Identi.ca handle: @kidehen
Google+ Profile: https://plus.google.com/112399767740508618350/about
LinkedIn Profile: http://www.linkedin.com/in/kidehen