W3C home > Mailing lists > Public > public-lod@w3.org > November 2010

class level mappings inferred from BTC2009 data

From: A.Nikolov <a.nikolov@open.ac.uk>
Date: Tue, 9 Nov 2010 00:53:02 +0000
To: "public-lod@w3.org" <public-lod@w3.org>
Message-ID: <2915005DADAFCB4899898B66C031AE9F5B7658C2B9@SALCEYCMS1.open.ac.uk>
Hi Andreas, Chris, and everybody,

Yesterday at the Consuming Linked Data Workshop I was asked about making public the set of class-level mappings produced from the BTC2009 data, which I mentioned in my presentation.
I have put a dump to http://people.kmi.open.ac.uk/andriy/mappings-btc.nt for now, Chris, I can put those to R2R if it's ok.

In short, this is a set of class subsumption mappings (about 50000) produced from the overlapping instance sets in the Billion Triple Challenge 2009 dataset using a light-weight machine-learning method.
As i said, there is some noise in the set because 
(a) machine learning was not 100% accurate (precision around 90% measured using cross-validation on a training set) and 
(b) some info became obsolete since last year.

Best regards,
Andriy Nikolov

-- 
The Open University is incorporated by Royal Charter (RC 000391), an exempt charity in England & Wales and a charity registered in Scotland (SC 038302).
Received on Tuesday, 9 November 2010 00:53:40 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 16:21:06 UTC