Mírian Bruckschen of Pontifícia Universidade Católica do Rio Grande do Sul, and colleagues, will present a paper entitled Named Entity Recognition in the Legal Domain for Ontology Population (for the full text of the paper, click here for the conference proceedings in PDF and scroll down to the page numbered 16) at SPLeT 2010: The 3rd Workshop on Semantic Processing of Legal Texts, to be held 23 May 2010 in Malta.
Here is the abstract of the paper:
This paper presents the overall problem of privacy risk assessment in the software industry and the difficulty to deal with all normative sources that regulate privacy matters. This problem encompasses the hard task of representing all the relevant information and keep it updated. Ontologies are the main mechanism for domain-specific knowledge representation in the Semantic Web context, but their manual maintenance is expensive and error-prone. Following the ontology learning trend, this paper presents an approach to automatically populate a legal ontology from legal texts through the Named Entity Recognition task and an experiment on this approach. Legal ontologies have been an active topic of research for quite a while, but on specific domains such as data privacy there is still a lack of such resources. The experiment described in this paper is run over a corpus of legal and normative documents for privacy, shows promising results and presents opportunities for the continuation of this research.