Towards an ontology of ElPub/SciX: a proposal

Abstract

A proposal is presented for a standard ontology language defined as ElPub/SciX Ontology, based on the content of a web digital library of conference proceedings. This content, i.e., ElPub/SciX documents, aims to provide access to papers presented at the total editions of the International Conference in Electronic Publishing (ElPub). After completing its 10th years in 2006, ElPub/SciX is now a comprehensive repository with over 400 papers. Previous work has been used as a basis to build up the ontology described here. It has been presented at Elpub2004 and it dealt with an Information Retrieval System using Computational Linguistics (SiRILiCo). ElPub/SciX ontology constitutes a lightweight ontology (classes and just some instances) and is the result of two basic procedures. The first one is a syntactic analysis carried out through the Syntactic Parser-VISL. This free tool, based on lingsoft's ENGCG parser, is made available through the Visual Interactive Syntactic Learning, a research and development project at the University of Southern Denmark, Institute of Language and Communication (ISK). The second one, carried out after that, is a semantic analysis (concept extraction) conducted through GeraOnto, an acronym that stands for “generating an ontology”, which extracts the concepts needed in order to build up the ontology. The program has been developed by Gottschalg-Duque, in 2005, in Brazil. The ensuing ontology is then edited via Protégé, a free, open source ontology editor. The motivation to carry out the work reported here came from problems faced during the preparation of a paper to Elpub2006, which aimed to present data about a number of aspects regarding the ElPub/SciX collection. While searching the collection, problems with the lack of standardization of authors and institutions names and the non-existence of any control of keywords had been identified. Such problems seem to be related to an apparent absence of “paper preparation” before entering into the SciX database. Lack of preparation, in turn, has brought about the desire of finding a solution, which is expected to support the work of those interested in searching the collection to retrieve information. ElPub/SciX ontology, therefore, is seen as that helping solution to support ElPub information retrieval.