Discovering the Topical Evolution of the Digital Library Evaluation Community

Abstract

The successful management of textual information is a rising challenge for all the researchers’ communities, in order firstly to assess its current and previous statuses and secondly to enrich the level of their metadata description. The huge amount of unstructured data that is produced has consequently populated text mining techniques for its interpretation, selection and metadata enrichment opportunities that provides. Scientific production regarding Digital Libraries (DLs) evaluation has been grown in size and has broaden the scope of coverage as it consists a complex and multidimensional field. The current study proposes a probabilistic topic modeling implemented on a domain corpus from the JCDL, ECDL/TDPL and ICADL conferences proceedings in the period 2001-2013, aiming at the unveiling of its topics and subject temporal analysis, for exploiting and extracting semantic metadata from large corpora in an automatic way.