UNIVERSITY OF HERTFORDSHIRE
             COMPUTER SCIENCE RESEARCH COLLOQUIUM

                            presents

     "Applying Data Mining Techniques to Sentiment Analysis 
                     for MediaEvaluation"


                       Dr. Daoud Clarke
       (Computer Science, University of Hertfordshire, UK)

                    23 March 2011  (Wednesday)
                      Meeting Room LD454

                  Hatfield, College Lane Campus
                            1 -2  pm

                  Everyone is Welcome to Attend

                  Refreshments will be available

Abstract:


This talk describes work from an ongoing Knowledge Transfer 
Partnership (KTP) project, between the University of Hertfordshire 
and Metrica, a company based in London.

Metrica is a media analysis company recognised as a world leader in
media analysis, whose clients include some of the top technology
companies, charities and government agencies. Their strengths are in
manual analysis of traditional (print) media.  However the explosion in
social media, and its corresponding increase in importance, has led them
to explore methods of automating this analysis, since there are too many
documents to analyse manually.

In this talk, we discuss our attempts to apply sentiment analysis to
Metrica's client databases, both in traditional and social media.  We
describe the data-mining and linguistic techniques used, the problems
that arose, and our current progress towards solving them.   In
particular, we will talk about the problem of unbalanced data, where
most of the input is in the 'uninteresting' class; the construction of
different kinds of representation of the documents; the process of
identifying effective classifiers; and the use of active learning, to
reduce the required amount of manually classified data.

---------------------------------------------------
Hertfordshire Computer Science Research Colloquium
http://homepages.stca.herts.ac.uk/~nehaniv/colloq