MANENT: An infrastructure for integrating, structuring and searching digital libraries

Abstract

Digital Libraries represent the commitment of research communities to preserve authoritative and well structured sources of knowledge, and to share archival organisations, methods and resources thanks to systems relying on standard metadata formats. This chapter describes some natural language processing techniques exploited for automatically extracting structural information from documents stored in Digital Libraries, based on the exposed metadata. The most prominent results achieved in this area are surveyed and discussed. As an example of an infrastructure for integrating, structuring and searching Digital Libraries based on natural language processing and semantic web techniques, we discuss the MANENT system. MANENT is a working prototype offering services of Digital Library content management and record classification and retrieval. It is hosted on a server at the Computer Science Department of Genova University and, starting from 2011, it will become publicly available. 475,000 records drawn from 138 repositories that all over the world expose OAI-PMH services have been downloaded, stored, and their automatic classification is under way. © 2011 Springer-Verlag Berlin Heidelberg.