From user needs to expert knowledge: Mapping laymen queries with ontologies in the domain of consumer mediation

Abstract

We believe that an important element towards the improvement of online legal services lies in the ability to identify legally relevant textual segments in users queries. In this line, this paper presents a case study in the processing of citizens queries in the domain of consumer justice from the point of view of legal domain semantics. The analysis is made in the framework of the ONTOMEDIA project, which aims at the design of a semantic platform enabling users and professional mediators to meet in a community-driven Web portal. The paper first presents the characteristics of the platform and its requirements; then reports the term extraction methodology from a corpus of consumer queries, and draws some conclusions on the double nature of laymen representation of legal problems: both terminological-conceptual and script-like (in the form of storytelling narratives). The paper further proposes an annotation structure based on these dimensions, which can be exploited to train machine learning algorithms to automatically recognize legally relevant text fragments. The paper concludes with the discussion of open issues for future work. © 2012 Springer-Verlag.