[hcls] Extracting conclusion sections from PubMed abstracts and turning them into linked data

I did some preliminary work making use of conclusion sections in abstracts, 
representing them using Semantic Web / Linked Data technologies and aligning 
it with existing datasets produced by the HCLS IG. The e-mail became so long 
that I decided to turn it into a pre-print publication.

Direct link to PDF:
http://precedings.nature.com/documents/3775/version/1/files/npre20093775-1.pdf

======
Extracting conclusion sections from PubMed abstracts for rapid key assertion 
integration in biomedical research

Available from Nature Precedings
<http://precedings.nature.com/documents/3775/version/1>

Key assertions are extracted from "conclusions" sections of PubMed abstracts
and converted into Semantic Web / Linked Data format. The results are made
accessible via  files, a SPARQL endpoint, and a  faceted search interface.
Conclusion sections are identified as valuable resources for
machine-augmented key assertion identification and integration in the
biomedical domain.  Results are discussed and opportunities for future work
and cooperation are highlighted.
======

If you just want to have a quick look at the outcome without bothering 
reading the paper, then:
1) have a look at http://hcls.deri.org/atag/explorer/
(works in every browser except IE. This is just a rough prototype, it still 
does not link back to source / evidence etc.).
2) do a search for "Schizophrenia"
3) Further narrow down search results by entering additional search terms or 
selecting facet values on the left.

I'm happy about feedback. I look forward to integrate this with other 
ongoing work in the HCLS IG. Together with Anja I will also look into making 
data from the LODD task available in the search interface above (a large 
part of SIDER has already been integrated, for example).

Cheers,
Matthias Samwald

DERI Galway, Ireland
http://deri.ie/

Konrad Lorenz Institute for Evolution & Cognition Research, Austria
http://kli.ac.at/ 

Received on Thursday, 17 September 2009 13:08:35 UTC