Joint topic and perspective model, lin 2008

From Cohen Courses
Jump to navigationJump to search

Citation

W.-H. Lin, E. Xing, and A. Hauptmann. A joint topic and perspective model for ideological discourse. In ECML PKDD ’08: Proceedings of the European conference on Machine Learning and Knowledge Discovery in Databases - Part II, pages 17–32, Berlin, Heidelberg, 2008. Springer-Verlag.

Abstract from paper

Polarizing discussions on political and social issues are common in mass and user-generated media. However, computer-based understanding of ideological discourse has been considered too difficult to undertake. In this paper we propose a statistical model for ideology discourse. By ideology we mean “a set of general beliefs socially shared by a group of people.” For example, Democratic and Republican are two major political ideologies in the United States. The proposed model captures lexical variations due to an ideological text’s topic and due to an author or speaker’s ideological perspective. To cope with the non-conjugacy of the logistic-normal prior we derive a variational inference algorithm for the model. We evaluate the proposed model on synthetic data as well as a written and a spoken political discourse. Experimental results strongly support that ideological perspectives are reflected in lexical variations.

Information

This paper presents a model for modeling perspectives in text. It assigns every word a topical weight indicating how often it was chosen depending on the topic, and an ideological weight which depends on the perspective of the speaker or author. However, it does not distinguish between different topics, but assumes all text is about the same topic. One of the datasets used is the Bitterlemons dataset.

External link: [1]