Abstract
In this paper, we observe the eects that discourse function
attribute to the task of training learned classiers for sentiment analysis.
Experimental results from our study show that training on a corpus of
primarily persuasive documents can have a negative eect on the performance
of supervised sentiment classication. In addition we demonstrate
that through use of the Multinomial Nave Bayes classier we can
minimise the detrimental eects of discourse function during sentiment
analysis.
Original language | English |
---|---|
Title of host publication | Computational Linguistics and Intelligent Text Processing: 15th International Conference, CICLing 2014, Kathmandu, Nepal, April 6-12, 2014, Proceedings |
Editors | Alexander Gelbukh |
Place of Publication | Berlin |
Publisher | Springer Verlag |
Pages | 45-52 |
Volume | II |
ISBN (Print) | 978-3-642-54903-8, 978-3-642-54902-1 |
DOIs | |
Publication status | Published - 2014 |