Domain Analysis for Knowledge Organization (eBook)

Tools for Ontology Extraction

Richard Smiraglia (Autor)

eBook Download: PDF | EPUB

2015 | 1. Auflage
116 Seiten
Elsevier Science (Verlag)
978-0-08-100188-2 (ISBN)

Infos

Domain analysis is the process of studying the actions, knowledge production, knowledge dissemination, and knowledge-base of a community of commonality, such as an academic discipline or a professional community. The products of domain analysis range from controlled vocabularies and other knowledge organization systems, to scientific evidence about the growth and sharing of knowledge and the evolution of communities of discourse and practice.In the field of knowledge organization- both the science and the practice­ domain analysis is the basic research method for identifying the concepts that will be critical building blocks for knowledge organization systems. This book will survey the theoretical rationale for domain analysis, present tutorials in the specific methods of domain analysis, especially with regard to tools for visualizing knowledge domains. - Focuses on the science and practice of organizing knowledge - Includes step-by-step instructions to enable the book to be used as a textbook or a manual for researchers

Richard Smiraglia is a Professor in the Knowledge Organization Research Group at the iSchool, University of Wisconsin, Milwaukee, USA. He has defined the meaning of 'a work' empirically, and has revealed the ubiquitous phenomenon of instantiation among information objects. Recent work includes empirical analysis of social classification, and epistemological analysis of the role of authorship in bibliographic tradition. His 'Idea Collider'' research team is working on a unified theory of knowledge. An Associate Researcher of the eHumanities Group, Amsterdam, he is a collaborating member of the Knowledge Space Lab effort to map the evolution of knowledge in Wikipedia. He holds a PhD (1992) from the University of Chicago. He is editor-in-chief of the journal Knowledge Organization.

Textauszug

2

Domain analysis as a methodological paradigm in knowledge organization

Abstract

In domain analysis, the purpose is to reveal the contours of held knowledge, whether that be in the form of live discourse or recorded documentation, by analyzing the elements of specific communities who share a common ontology, or knowledge base. The objectives of domain analysis are to map and visualize the intellectual parameters of shared knowledge in a given community, such that results can be put to use in knowledge organization systems for the furtherance of the community’s own discourse and for its intellectual contributions at large. The evolution of empirical methods can be observed in the literature of the domain, which is found predominantly in the proceedings of international ISKO conferences and in the journal Knowledge Organization. Nearly a 100 studies have been conducted and reported in the domain’s formal literature and summarized here. The majority of analytical studies use empirical methods such as metric or terminological techniques, but large numbers of discourse analyses, genre analyses, and epistemological analyses also are attempted. Fewer critical studies and historical analyses have been generated but some are represented. A wide variety of domains has been analyzed in these studies. Approximately 50 domains have been studied 1–4, and 22 studies reported domain analysis of aspects of knowledge organization.

Keywords

Held knowledge

Discourse

Recorded documentation

Common ontology

Knowledge bases

2.1 A methodological paradigm in KO

Domain analysis is one methodological paradigm within the science of knowledge organization. In domain analysis, the purpose is to reveal the contours of held knowledge, whether that be in the form of live discourse or recorded documentation, by analyzing the elements of specific communities who share a common ontology, or knowledge base. The objectives of domain analysis are to map and visualize the intellectual parameters of shared knowledge in a given community, such that results can be put to use in knowledge organization systems for the furtherance of the community’s own discourse and for its intellectual contributions at large. Put simply, our purpose is to visualize the concepts shared by a group that works together, to discern the working vocabulary that defines those concepts, and to map the relationships among the concepts. Domain analysis is one aspect of knowledge organization where both ontology and epistemology come together; we want to see which concepts constitute the knowledge base, or ontology, of discourse in a community, but we also want to see how those concepts are used and understood.

Domain analysis as a named, purposive methodological paradigm in knowledge organization can be dated from Hjørland’s catalyzing paper in 2002; although the roots go back further to the understanding that knowledge organization had come to a domain-centric postmodern viewpoint (Mai, 1999). Rather than seeking universally applicable rules for formulating a single ontology, postmodern KO seeks to reveal the ontological parameters of each specific community, choosing instead to search for techniques for interoperability to allow cross-domain communication. Domain analysis as a methodological paradigm, then, has two demands. The first is that each domain be analyzed exhaustively and continuously. The second is that to facilitate theory building, there must be replicated and continuous analysis both of specific domains and of their multiplicity of domains together. That is, it is imperative that knowledge organization as a science turn its metaphorical microscope to look at every possible domain from the workplace to the neighborhood to the household to the academic disciplines and beyond. As marching orders for a new science the turn to domain analysis has meant that more KO scholars are needed to analyze more domains. But it also has meant that more analyses of already analyzed domains also are required.

In the 12 years or so in which KO has turned its attention specifically in a domain-analytical direction, there has been a modest increase in the number and frequency of studies specifically devoted to using Hjørland’s 11 approaches to the analysis of domains for the purpose of revealing their shared ontologies. In this chapter, we look briefly at those studies, which appear in KO’s three principal venues. Interestingly, we know from research (Smiraglia, 2011, 2012, 2013a) that those venues are primarily the biennial international conference proceedings from the International Society for Knowledge Organization (published in the series Advances in Knowledge Organization) and the journal Knowledge Organization, and a smattering of papers from the information science literature at large. Other sources are doctoral dissertations produced in schools of KO, although most of these are eventually reported formally in one of the other three venues.

2.2 Domain-analytical literature from the KO domain

Domain analysis has been explicitly a critical aspect of the knowledge organization domain since Hjørland’s charge to the community in 2002. The evolution of empirical methods can be observed in the literature of the domain, which is found predominantly in the proceedings of international ISKO conferences and in the journal Knowledge Organization. In this section I review, briefly, studies from those venues as well as from some doctoral dissertations. The point of this curt review is to point readers to specific examples of the application of specific methods.

2.2.1 International ISKO conference proceedings

Forty-one domain-analytical papers from ISKO International conferences were sorted methodologically for this review. That is, each paper was categorized using 1 of the 11 approaches to domain analysis from Hjørland’s list. Interestingly, only 7 of the 11 approaches have been used in papers in the proceedings. Most often used are terminological or informetric studies, and discourse analyses. Obviously, bibliometrical studies constitute the most used forms of domain analysis in published knowledge organization research. One could hazard a guess that this is because the written and indexed record is so plentiful. But it is important to note that many aspects of knowledge organization and the domains studied are not indexed by the most prominent indexing services. This means that much of the work must be done manually, which is time and labor intensive, but also involves much cleaning of data, usually in the form of poorly or inconsistently formatted citations. Interestingly 33 definable domains have been analyzed in ISKO conference papers since 2002, ranging from archival science to transport engineering. Only two domains have been visited more than once; there were two studies of gender studies, and eight studies of the aspects of knowledge organization. The domain-analytical lens has been turned inward more often than outward it would seem. Papers in these clusters are reported here and reports within each cluster are chronological.

2.2.1.1 Producing literature guides or subject gateways

Coleman (2004) used content maps and pathfinder networks to test the domain vocabulary of engineering novices. Madalli et al. (2014) drew on a variety of known KOSs for music to generate a faceted ontological representation.

2.2.1.2 Empirical user studies

Campbell (2004) conducted a qualitative study of gay and lesbian information users to detect the value of facet analysis for generating flexible and adapatable approaches to browsing systems. Pajarillo (2006) conducted an ethnographic study of home health care nurses, using diaries, participant observation, focus groups, and questionnaires. Loehrlein (2008) used content analysis of a corpus of records management texts to discover Wittgensteinian form-of-life contexts. Participants subsequently ranked interpretations of the texts based on complexity judgments. Sanatjoo (2010) used a mixed-methods work-task-oriented methodology (WOM) to develop a thesaurus for a plant pathology academic unit. Actors, tasks, and information behaviors were identified, conceptualized, and subjected to facet analysis. This methodology is similar to, but not quite coextensive with Cognitive Work Analysis (CWA). Orrico et al. (2012) used a case study of researchers in transport engineering together with content analyses of artists' manifestos to extract a theoretical model of the use of metaphors. Frâncu and Popescu (2014) combined content analysis and interviews with traditional bibliometric measures to develop a profile of the cultural shift in the evolution of Romanian knowledge organization resulting from a 20-year period during which applied research continued in libraries while theoretical work was discontinued in academic institutions.

2.2.1.3 Bibliometrical studies

López-Huertas (2006) studied Spanish-language gender studies in Uruguayan publications selecting terms from titles, abstracts, and main headings in monographs, articles, and reports, and section or column headings in specialized periodicals. Chaomei et al. (2008) used bibliographic records from the Web of Science™ (WoS)™ to look for culture and geographic identity in the domain of astronomy surrounding the Sloan Digital Sky Survey. Three text-mining systems were used for the...

Produkt-Details

• Erscheint lt. Verlag: 15.6.2015
• Sprache: englisch
• Themenwelt: Geisteswissenschaften ► Sprach- / Literaturwissenschaft
• Themenwelt: Mathematik / Informatik ► Informatik ► Software Entwicklung
• Themenwelt: Sozialwissenschaften ► Kommunikation / Medien ► Buchhandel / Bibliothekswesen
• Themenwelt: Wirtschaft ► Betriebswirtschaft / Management ► Unternehmensführung / Management
• Themenwelt: Wirtschaft ► Betriebswirtschaft / Management ► Wirtschaftsinformatik
• ISBN-10: 0-08-100188-6 / 0081001886
• ISBN-13: 978-0-08-100188-2 / 9780081001882