Abstract

There are several methods for constructing an ontology. Among the automatic methods, one approach is the extraction of terms from domain documents and their subsequent extraction. In this case, the first step of the process is the extraction of noun phrases that are potential candidates to be components of the terminology of the area of interest. This article describes an automatic tool for the Brazilian Portuguese language that extracts noun phrases that can be adopted as terms for a certain domain. In addition, the system couples the extracted terms into a top-level ontology, which results in an initial ontology that can be further refined. To couple with the ontology an anchor term was used, and a statistic analysis showed that the use of the term anchor leads to an improvement in the performance of the system. The tool described in this article was used to select terms to be used in an ontology for the power sector domain. Also, the precision in the creation of the ontology was evaluated. The technique was able to generate the correct hierarchy for 70\% of the terms.

Links and resources

Tags