Structured argumentation modeling and extraction: Understanding the semantics of parliamentary content

Abstract

In information overload scenarios, the volume, structure and complexity of generated data represent a challenge that hinders the content comprehension. Aiming to overcome these dissuasive issues, the modeling and extraction of arguments in textual content has become a prominent topic in the information retrieval field. In this paper, we propose a new argumentation model, where different semantic components and their relationships are considered. Our proposal aims to enhance state of the art approaches, which limit their scope to identifying chunks of text as argumentative or not, leading to large amounts of texts left unanalyzed. The presented model, differently to domain-specific corpus methods, is designed to enable a generic, cross-lingual semantic annotation that promotes reusability. As a proof of concept, the model is exemplified in a case study for an e-government platform intended to annotate semantically, and provide information retrieval and filtering functionalities on content produced in the Spanish Parliament