Annotating a Corpus for Building a Domain-specific Knowledge Base

Proceedings of the Fourth International Conference on Language Resources and Evaluation (LREC 2004)

Abstract

The project described in this paper seeks to develop a knowledge base for the domain of data processing in construction - a sub-domain of mechanical engineering - based on a corpus of authentic natural language text. Central in this undertaking is the annotation of the relevant linguistic and conceptual units and structures which are to form the basis of the knowledge base. This paper describes the levels of annotation and the ontology on which the knowledge base is going to be modelled and sketches some of the linguistic relations which are used in building the knowledge base.