Home // ALLDATA 2017, The Third International Conference on Big Data, Small Data, Linked Data and Open Data // View article

Topic Models to Contextualize and Enhance Text-Based Discourses Using Ontologies

Authors:
Dimitris Gkoumas
Réka Vas

Keywords: Discourse contextualization; discourse enhancement; clustering topic model; probabilistic topic model; ontologies in NLP.

Abstract:
Public policy-making has a clear and unique purpose: achieve a desired goal that supports the best interest of all members of the society by providing guidance for addressing selected public concerns. Examples include clean air, healthcare, waste management etc. The identification of social targets and pathways – by which these targets could be reached – are at the core of policy-making. This paper is part of an ongoing research aiming at enhancing public policy-making in the field of waste management by contextualizing and enriching text-based, Web forum discourses on waste management. For that purpose, an ontology model describing the waste management domain has been created. In the next step, the actual forum discussions are connected to one or more subdomains of the ontology by determining what proportion of the sub-domain is covered by that discourse. Finally, applying text mining techniques semantically enriched domain concepts are assigned to the discourse. This paper also provides a critical discussion on two text mining approaches that could be applied for this purpose, also highlighting points that deserve further investigation.

Pages: 78 to 82

Copyright: Copyright (c) IARIA, 2017

Publication date: April 23, 2017

Published in: conference

ISSN: 2519-8386

ISBN: 978-1-61208-552-4

Location: Venice, Italy

Dates: from April 23, 2017 to April 27, 2017