Enhancing document modeling by means of open topic models

Abstract

We present a topic classification model using the Dewey Decimal Classification (DDC) as the target scheme. This is done by exploring metadata as provided by the Open Archives Initiative (OAI) to derive document snippets as minimal document representations. The reason is to reduce the effort of document processing in digital libraries. Further, we perform feature selection and extension by means of social ontologies and related web-based lexical resources. This is done to provide reliable topic-related classifications while circumventing the problem of data sparseness. Finally, we evaluate our model by means of two language-specific corpora. This paper bridges digital libraries on the one hand and computational linguistics on the other. The aim is to make accessible computational linguistic methods to provide thematic classifications in digital libraries based on closed topic models as the DDC.

Lefa la Tsebo: A Proposed Model for Framing Tshwane University of Technology’s Libraries Policies

“…The authors draw on a focused literature review across various disciplines and empirical evidence collected from a purposive sample of 40 participants in a case study by Molepo (2018). The qualitative evaluation method is used to evaluate a set of models: models for academic library activities (see Baglier and Caswel 2016;Dongardive 2013;Joint 2011;Mehler and Waltinger 2009;Sigwald 2016), models for higher education (see Hay and Van Gensen 2008;Van Zijl , Gericke, and Machet 2006;White 2017), as well as models depicting changes in society's knowledge system (see Etzkowitz and Leydesdorff 1995;1997;2000;Funtowicz and Ravetz 1993;Gibbons et al 1994;Lundvall 2005;Schafer 1983;Slaughter and Rhoades 2004), and Dick's (1982) model of society's knowledge system, which builds on the earlier work of Machlup (1962), Holzner and Marx (1979) and others. A self-developed set of criteria is used as a tool to test the models towards a selection of elements and features that are then used to table the proposed model for TUT's libraries.…”