Mining digital library evaluation patterns using a domain ontology

Abstract

Scientific literature is vast and therefore the researchers need knowledge organization systems to index, semantically annotate and correlate their bibliographic sources. Additionally they need methods and tools to discover scientific trends and commonly acceptable practices or areas for further investigation. This paper proposes a clustering-based data mining process to identify research patterns in the digital libraries evaluation domain. The papers published in the proceedings of a well known international conference in the decade 2001-2010 were semantically annotated using the Digital Library Evaluation Ontology (DiLEO). The generated annotations were clustered to portray common evaluation practices. The findings highlight the expressive nature of DiLEO and underline the potential of clustering in the research activities profiling.