The 8th European Networked Knowledge Organization Systems/Services (NKOS) Workshop, organized by Traugott Koch, Douglas Tudhope, Marianne Lykke Nielsen and Koraljka Golub, took place on October 1, 2009, in Corfu, Greece, as part of ECDL 2009. The full-day workshop was attended by 28 colleagues from 14 countries on 3 continents. All presentations and abstracts are available on the workshop website <http://www.comp.glam.ac.uk/pages/research/hypermedia/nkos/nkos2009/programme.html>.
NKOS is a community of researchers, developers and practitioners seeking to enable knowledge organization systems (KOS) (e.g., classifications, gazetteers, lexical databases, ontologies, taxonomies and thesauri) as networked interactive terminology and knowledge organization services via the Internet.
This year, the workshop covered rather diverse topics, spread widely over the list of themes in the Call for Papers [cf. <http://www.comp.glam.ac.uk/pages/research/hypermedia/nkos/nkos2009/>]. Even the new topic, "Linked data on the web, SKOS based tools and services," was represented. For the first time, we invited formal posters as a main part of the workshop and devoted an extended presentation and discussion time, both bilaterally with the author and in the plenary. Throughout the workshop, a maximum of discussion time seemed, as usual, not sufficient for all the questions and issues, a situation emanating from the good mix between high level experts and interested colleagues and practitioners.
As in previous years, participants had the opportunity for further discussion and planning of active collaborations over dinner after the workshop, taking advantage of the mutual introductions of all participants and their areas of interest during the event itself. Four of the ten presentations/posters were carried out by colleagues who have presented before at NKOS workshops, and a third of the participants were returning from earlier years. Among related communities, Information Retrieval, Semantic Web and ISKO (International Society for Knowledge Organization) were well represented.
"Design models and methodologies, including user-centred approaches", was the theme of the first (and most extensive) session, which was chaired by Stella Dextre Clarke.
Claudio Gnoli, University of Pavia, Italy, reported from a working group, establishing the requirements of an updated section on philosophy in the Universal Decimal Classification (UDC), based on the principles of facet analysis. Relevant facets, enabling a more logical structure, are, e.g., schools, periods, applications, methods, and viewpoints. Notational solutions need to be optimized for automatic parsing, appropriate nesting of facets and reduced complexity of notations. As presented by Claudio in last year's NKOS workshop, information retrieval and browsing need to be improved through a unique notation for each concept, explicit relationships and an unambiguous syntax.
UDC provided the context for Aida Slavic's (UDC Consortium) presentation too. Management issues in networked classification schemes were discussed, starting from the problem of how to represent them properly in SKOS encoding. The management of persistent identifiers for classes/concepts and the versioning, considering frequent changes and pre-combined concepts in a classification system, are experienced as far from trivial problems. It is important also to consider workflows for managing established systems and to work around current limitations in SKOS.
A Vocabulary Bank for Education as part of the UK ASPECT project was described and demonstrated by Rob Tice, a registry in which locally stored (multilingual) terms, controlled vocabularies, especially thesauri, metadata schemas, application profiles and services, can be managed, documented and "published". Both schemas and individual terms feature descriptive records, mapped behind the scenes to DC as a common format and accessible for real time searching. Among the exchange formats on offer are Zthes, VDEX and SKOS Core.
User interface design issues for services offering search-term recommendation and interactive query expansion were discussed by Philipp Schaer, GESIS Social Science Information Center, Germany. He demonstrated a prototype for such a recommender tool, integrated into the sowiport portal. It uses a probabilistic latent semantic analysis and a support vector machine-based classification tool, mapping free user terms to the controlled vocabulary of the social science literature database SOLIS. Early evaluations were related to issues such as the number of actions required by the user, the display of intermediate results in a separate or the same screen, the phase when suggestions are welcome and the selection of defaults.
The poster sessions, again chaired by Stella Dextre Clarke, featured three posters with extended discussion time.
Uma Murthy, from Virginia Tech, is developing a Crisis, Tragedy, and Recovery ontology for information integration in CTRnet, an integrated distributed digital library network. It will be used for browsing and query expansion as well. The vocabulary is drawn from scientific literature, political documents, professional focus groups and user interactions, each using very different terminology. The ontology will need to be evaluated in relation to the coverage of the content, coherence, consistency and the hospitality to new content.
The second poster, presented by Carmen Reverté Reverté (Aquatic Ecosystems IRTA, Spain), dealt with KOS mapping. In a European Aquatic Sciences subject gateway project, she evaluates two different strategies to advance to a common knowledge organization structure: a) mapping heterogeneous KOS in the field, using SKOS and OWL with related tools and b) creating a new multidisciplinary thesaurus for the purpose.
Due to special circumstances, the poster on the UK project EASTER could not be presented by the authors. Automated subject metadata generation tools will be evaluated and compared to manual cataloguing by experts, in order to find out how they might improve cataloguing and end-user searching, employing both controlled vocabulary and free terms.
"Semantic web applications and implications of KOS" was the first of two afternoon sessions chaired by Traugott Koch.
The work to encode the German bilingual STW Thesaurus for Economics with SKOS and to provide it as a web service was presented by Joachim Neubert, German National Library of Economics (ZBW). The SKOS representation has been published as "Linked Data" with a SPARQL endpoint, free for any non-commercial usage. A REST-based web service is in operation. Retrieval is intended to be supported in the form of query expansion, for both thesaurus-indexed structured data and fulltext. Users depositing scientific articles in the DSpace repository EconStor will be assisted in their indexing/tagging, as will professional cataloguers with an autosuggest service for thesaurus terms, similar to the one in use on the STW website.
Due to Visa-related problems, Azade Sanjari, working for the Encyclopedia of Iranian Architectural History, was not able to present. A special ontology and a Dublin Core Application Profile is used to integrate distributed digital repositories in their project. Harvested metadata is interpreted by the ontology and presented in a "semantic portal". The slides are available at the workshop website.
The last session gathered various topics from project contexts that have been discussed at earlier NKOS workshops.
The PERTAINS project, presented by Caroline Williams and Andy Priest from Intute and Manchester University, is creating tag recommender demonstrators for the UK national services COPAC, a research library OPAC aggregator, and the Intute database of scholarly web resources (already rich in subject metadata). It extends the developments of the EnTag project's DDC-based recommender to a much larger scale and supports personalization features of the services (personal collections). Tag suggestions will come from both SKOS-encoded controlled vocabularies via a RESTful web service and from unstructured sets of previous tags.
Philipp Mayr, GESIS Social Science Information Center, Germany, followed up on his paper at the NKOS workshop 2007 and presented the outcome of comprehensive evaluation efforts regarding the information retrieval effects of a net of mapped vocabularies (project KoMoHe). Intellectually produced cross-concordances between 28 controlled vocabularies in several languages (predominantly German) and with focus on Social Sciences have been used for searches in rather large databases to evaluate the retrieval effects of the mappings. Somewhat surprisingly, interdisciplinary searches show a much higher improvement in precision and recall than searches inside the same discipline. The greater amount of identical terms and of overlap in intradisciplinary vocabularies and databases is seen as a major reason. Thus, automatic mappings could be applied here in the future.
On a regular basis, European NKOS workshops have been following the development of standards in structured vocabularies, especially the thesaurus standard ISO 25964 and its predecessors. Stella Dextre Clarke reported that Part 1 "Thesauri for information retrieval" will be issued as a Draft International Standard before the end of this year and that work on Part 2 "Interoperability with other vocabularies" has begun. Improvements over its predecessors, ISO 2788 and ISO 5964, are, e.g., a clearer distinction between concept and term, more guidance on facet analysis and on thesaurus management software, completely revised multilingual examples and a data model based on extending the model in BS 8723. The case of a "non-symmetrical multilingual thesaurus" was especially discussed.
For the second year in a row, the US NKOS workshop series was continued in cooperation with CENDI (an interagency working group of senior scientific and technical information managers from 13 U.S. federal agencies), on October 22, 2009. It discusses approaches to the management of knowledge organization systems that will support emerging and future requirements and use cases.
The main forum for all related activities is the international NKOS discussion list. Information about events and publications, how to subscribe to the list and to share the work is available at the international NKOS website <http://nkos.slis.kent.edu/>.
Further information on this 8th European NKOS workshop along with links to all presentations can be found on the workshop website at <http://www.comp.glam.ac.uk/pages/research/hypermedia/nkos/nkos2009/programme.html>.
Copyright © 2009 Traugott Koch