Automatic indexing and ontologies: the consistency of research chronology and authoring in the context of Information Science

Abstract

Considering the large digital collections, as currently available on the web, have made unfeasible if one uses manual indexing, one alternative is automatic indexing. Information retrieval systems have advanced but they are still not efficient in terms of searching that are based in themes or concepts. One alternative to minimize the semantic gap in those systems is to make use of ontologies. Given that automatic indexing and ontologies are subjects of study with a long history in the field of Information Science (IS), we intend to identify and to analyze the consistency of studies between those subjects, specifically: i) to identify papers in the Information Science & Technology Abstracts (ISTA), Library & Information Science Source (LISS) and Library and Information Science & Technology Abstracts (LISTA) databases dealing with these topics, exploring their temporal distribution and affiliations of their authors; (ii) to verify the overlap of the retrieved articles in the referred databases and; (iii) to assess the relationship between the studies by mapping the citations in the corpora. We use as methodology an exploratory / descriptive study based on a systematic review. Based on the preliminary results we find that: the number of papers addressing the relationship between the two concepts does not exceed a few dozen; a very close percentage of researchers associated with the IS field along with those associated to Computer Science; a degree of coverage overlap ranged from 45% to 65%; and the existence of links between the numerous authors and their respective works apparently dissociated from each other.