From: Lin, Jimmy (NIH/NLM/NCBI) [E]
Sent: Tuesday, December 09, 2008 3:57 PM
To: NLM/NCBI List ncbi-seminar
Cc: jimmylin@umd.edu
Subject: Seminar: Tues 12/6,2pm: Is searching full text more effective
than searching abstracts?(Jimmy Lin)

--------------------
SEMINAR ANNOUNCEMENT
--------------------

Is searching full text more effective than searching abstracts?
Jimmy Lin (University of Maryland/NCBI)

When: Tuesday, December 16, 2008, 2pm
Where: 38A B2 NCBI Library

With the growing availability of full-text articles online, scientists
and other consumers of the life sciences literature now have the
ability to go beyond searching bibliographic records (title, abstract,
metadata) to directly access full-text content.  Motivated by this
emerging trend, I posed the following question: Is searching full text
more effective than searching abstracts?  This question is answered by
comparing text retrieval algorithms on abstracts and full text using a
collection of 162k articles from the TREC 2007 genomics track
evaluation.

In this talk, I focus on two aspects of full-text retrieval:
effectiveness and efficiency.  Experiments illustrate the value of
full text, particularly if retrieval is performed at the level of
paragraphs.  These results suggest that effective retrieval algorithms
should perform finer-grained analysis of full-text articles.

The benefits of full-text come at considerable cost in terms of the
increased volume of data that need to be processed.  To this end,
Google's MapReduce programming model provides a convenient framework
for taking advantage of multiple machines in a cluster.  I'll discuss
experiments that focus on the practicality of scaling to
ever-increasing datasets.