The multilingual thesaurus of LAURIN

Among the wide range of digital libraries, an interesting, yet quite neglected, subclass is constituted by those exclusively dealing with newspaper clippings. Compared with book-oriented digital libraries, clipping libraries are more difficult to seize, since they are wide and unstructured, and the subjects and content of a clipping are completely heterogeneous. LAURIN is an EU-funded project involving seventeen participants from several countries, including two software companies and a large group of libraries, whose main purpose is to set up a network of digitalized newspaper clipping archives that can be easily accessed through the Internet, for searching and retrieving clippings. The project also provides the libraries with models and methodologies to be used for scanning, digitalizing, storing, indexing, and making accessible newspaper clippings. The core of the LAURIN system is an integrated multilingual Thesaurus. In this paper we illustrate how to express the thesaurus in terms of a knowledge base, and how to exploit such a knowledge base to improve the fundamental task of clipping retrieval. Copyright 2002 ACM.