Extracting Ontology Hierarchies From Text

Ontologies are an approach for knowledge representation capable of expressing a set of entities and their relationships, constraints, axioms and vocabulary of a given domain. Manual construction of ontologies by domain experts and knowledge engineers is an expensive and time consuming task so, automatic and/or semi-automatic approaches are needed. Ontology Learning looks for automatically or semi-automatically identifying ontology elements like classes, taxonomic and nontaxonomic relationships, properties and axioms from textual resources. This article proposes a process for the automatic extraction of ontology taxonomic relationships from English texts using natural language processing techniques. Some experiments using a legal corpus were conducted in order to evaluate it. Initial results are promising.