Context Selection in a Heterogeneous Legal Ontology

Abstract

Ontology building in the legal domain is subject to ongoing research. Taxonomic ontologies provide for instance concept hierarchies for term definitions, annotations, query expansion and support for inferences. However, the context-dependent application of statuatory legal texts is hard to model, often leading to a limited ontology scope and fixed terminology to avoid conflicts. In previous work, we presented a method to create a lightweight heterogeneous ontology from textbooks offering connections between laws, while avoiding an error-prone and costly ontology alignment step. In our ontology, laws are linked by common contexts. We propose a new data model, so that the context can be explored and selected by a user, which is necessary for many applications, such as recommender systems. To obtain the relevant user context, we added a mechanism to retrieve linked laws from our ontology, given a scope of user interest and context information for each law.