Ontology-Based Information Extraction under a Bootstrapping Approach
Elias Iosif (National Center for Scientific Research (NCSR) “Demokritos”, Greece), Georgios Petasis (National Center for Scientific Research (NCSR) “Demokritos”, Greece) and Vangelis Karkaletsis (National Center for Scientific Research (NCSR) “Demokritos”, Greece)
Copyright: © 2012
The authors present an ontology-based information extraction process, which operates in a bootstrapping framework. The novelty of this approach lies in the continuous semantics extraction from textual content in order to evolve the underlying ontology, while the evolved ontology enhances in turn the information extraction mechanism. This process was implemented in the context of the R&D project BOEMIE1. The BOEMIE system was evaluated on the athletics domain.
Ontologies in OBIE systems provide the domain knowledge model required for the systems’ operation. This model can be a rather poor one (e.g., a flat list of athlete names, location names, etc., the so-called gazetteer lists) or a rich one (e.g., a model built using an ontology language like OWL, which enables the representation of complex entities or events as well as the reasoning over them) enabling the categorization of IE systems according to the level of ontological knowledge they use. In order to classify OBIE systems we follow the classification proposed in Nedellec (2006), according to which four different levels of ontological knowledge can be exploited by an IE system.