Author:
Mare Koit
Affiliation:
University of Tartu, Estonia
Keyword(s):
Debate, Internet Comment, Opinion, Judge, Winner.
Related
Ontology
Subjects/Areas/Topics:
Applications
;
Artificial Intelligence
;
Knowledge Engineering and Ontology Development
;
Knowledge-Based Systems
;
Natural Language Processing
;
Pattern Recognition
;
Symbolic Systems
Abstract:
Together with an increasing role of online media in human communication it is necessary to perform automatic analysis of online texts. In this paper, we are studying dialogues formed by opinion articles and their comments on Internet. Such a dialogue can be considered as debate between two teams. One team connects the commentators with positive and another – with negative comments about the initial opinion, i.e. the commentators who respectively, support or reject the opinion presented in the source text. The members of both teams can in any time have the floor what is different as compared with conventional spoken debate. Internet users who spontaneously give marks +1 or -1 to the comments act as a board of ‘judges’. The winner is the team with a bigger total sum of marks. For every comment, we also assign a point in a mental space which we call communicative space. The values +1, 0 or -1 of the coordinates of communicative space make it possible to classify the comments not only as
positive and negative but also as polite and impolite, friendly and hostile, etc. The set of comments forms a collective opinion about the main agent of the source text which introduces a social aspect into the text analysis. The further aim of this preliminary study is the automatic analysis of such debates.
(More)