Proceedings of the International Congress on Interdisciplinarity in Social and Human Sciences
dIScourSE StructurE And contEnt AnALYSIS:
Rui Talhadas U. Algarve FCHS/INESC ID Lisboa ([email protected])
Nuno Mamede U. Lisboa IST/INESC ID Lisboa ([email protected])
Jorge Baptista U. Algarve FCHS/INESC ID Lisboa ([email protected])
ABStrAct Content analysis is a relevant tool for many human and social sciences, such as Psychology and Sociology, among others. The detection of the structure of the texts is a relevant step in determining how the major content elements are organized. Besides text segmentation into paragraphs, sentences, and clauses, the use of discourse connectors is a fundamental element for the structuring of a text. These connectors include conjunctions and conjunctive adverbs, and they make explicit the meaning relations between sentences forming a text. In this paper, we illustrate a method for capturing the major components of texts and their explicit organization. For evaluation, the method is applied to discourse parsing but it could also be applied to many tasks of content analysis. This interdisciplinary method bridges topics from linguistics and computational linguistics, with possible uses in several areas of social sciences, where content analysis and discourse structure may be relevant.
Keywords: Content Analysis, Text/Discourse Parsing, Discourse Connectors, Portuguese.