Stefano Menini will present is PhD research to the Data and Web Science Group (focus group on Natural Language Processing and Information Retrieval, headed by Prof. Ponzetto) at the Mannheim University on April 18, 2016.
Title: “Agreement and Disagreement: Analyzing Points of View in Political Speeches”.
In this talk, we propose a novel approach to recognize divergences and similarities in points of views between different authors or speakers when they deal with the same topic. We first used the transcriptions of the 1960 U.S. Presidential Campaign speeches as a case study to investigate how different features contribute to classify two statements as in agreement or disagreement. We then propose a model to detect the agreement without using external features but relying solely on surface and latent features contained in the snippets we want to compare, and we tested it on a large set of Debates form Debatepedia.
Keywords: Digital Humanities, Quantitative History, Natural Language Processing.