The paper suggests extensions to SKOS Core to make explicit where concepts in a knowledge organization system have changed from one version of the system to another.

Artificial Intelligence Lab, Department of MIS, University of Arizona / The rapid proliferation of textual and multimedia online databases, digital libraries, Internet servers, and intranet services has turned researchers' and practitioners' dream of creating an information-rich society into a nightmare of info-gluts. Many researchers believe that turning an info-glut into a useful digital library requires automated techniques for organizing and categorizing large-scale information. This paper presents research in which we sought to develop a scaleable textual classification and categorization system based on the Kohonen's self-organizing feature map (SOM) algorithm. In our paper, we show how self-organization can be used for automatic thesaurus generation. Our proposed data structure and algorithm took advantage of the sparsity of coordinates in the document input vectors and reduced the SOM computational complexity by several order of magnitude. The proposed Scaleable SOM (SSOM) algorithm makes large-scale textual categorization tasks a possibility. Algorithmic intuition and the mathematical foundation of our research are presented in detail. We also describe three benchmarking experiments to examine the algorithm's performance at various scales: classification of electronic meeting comments, Internet homepages, and the Compendex collection.

The marginalization of people through classification schemes results in inadequate access to information about these people when the context is, for example, a bibliographic classification system. When the context is the classification of the people themselves, they themselves are underrepresented, for instance, by society and government support. Taking the case of the natural disaster survivor, this paper explores appropriate steps to devising an accurate classification scheme of the survivors.

From the early information retrieval systems to recent web search engines, most systems ask user to express their information need in a query. While the online information systems, such as search engines and digital libraries, bring the great opportunity of accessing huge amount of information directly, they also present challenges on users’ competence of formulating good queries. In library, thesauri and classification schemes are not only indexing tools, but also concept consultation tools. Thesauri and classification schemes haven’t fully adapted in to digital library and IR system, because there are not universal concept hierarchies. This essay introduces the ideal of using query-based concept structures to improve subject access to digital libraries. It reports some early work of an ongoing project that explores the usability and effectiveness of query-based concept structures, which naturally connects static knowledge in the information collection and user dynamic information need, as intermediary. Methodologies and experimental designs are laid out. The preliminary results are presented and further research is discussed.

Artificial Intelligence Lab, Department of MIS, University of Arizona / The National Science Digital Library (NSDL), launched in December 2002, is emerging as a center of innovation in digital libraries as applied to education. As a part of this extensive project, the GetSmart system was created to apply knowledge management techniques in a learning environment. The design of the system is based on an analysis of learning theory and the information search process. Its key notion is the integration of search tools and curriculum support with concept mapping. More than 100 students at the University of Arizona and Virginia Tech used the system in the fall of 2002. A database of more than one thousand student-prepared concept maps has been collected with more than forty thousand relationships expressed in semantic, graphical, node-link representations. Preliminary analysis of the collected data is revealing interesting knowledge representation patterns.

James Duff Brown (1862-1914), an important figure in librarianship in late nineteenth and early twentieth century England, made contributions in many areas of his chosen field. His Subject Classification (SC), however, has not received much recognition for its theoretical and practical contributions to bibliographic classification theory and practice in the twentieth century. This paper discusses some of the elements of SC that both did and did not inform future bibliographic classification work, considers some contrasting evaluation methods in the light of advances in bibliographic classification theory and practice and of commentaries on SC, and suggests directions for further research.

INTRODUCTION Medical professionals seek to capture papers which can be located via keyword or free text search in digital libraries or on the web but are also interested in finding material that has not yet been indexed in on-line databases. Search engines provide a multitude of results [1]. Social bookmarking, where users tag items for their own use, offers a way to locate new and relevant information. CiteULike (citeulike.org), a social bookmarking service, allows articles to be tagged with useful keywords for later retrieval. RELATED STUDIES A previous study [2] compared social bookmarking to existing information organisation structures and found similarities in terminology use and intriguing differences. A sample of articles tagged on CiteULike was examined for contextual differences in keyword usage between users of social bookmarking sites, authors and indexers. Many tags were related to thesaurus terms (descriptors), but were not formally in the thesaurus. [2] This study examines how term usage patterns in tags, keywords and descriptors suggest a similar (or differing) context between users, authors and intermediaries. METHODOLOGY This study examines the use of tags on CiteULike from three medical or biology journals (JAMA, Proteins, and Journal of Molecular Biology) indexed in Pubmed. 1299 unique articles were retrieved from Citeulike; Medical Subject Headings (MeSH) were collected from Pubmed. Articles were analysed using standard informetric techniques to examine the use of user assigned tags and their Pubmed assigned MeSH index terms. Data was analysed for term usage and categorised to see what contextual clues users expose in their tag use. RESULTS Articles were tagged by up to 14 users (average 2-4). 1449 unique tags were used in the data set. Some articles were heavily tagged by users (max. 29, min. 1, median 2). Descriptors were more heavily assigned to articles (2746 unique descriptors). Articles had, on average, 10 descriptors assigned (max. 40, min. 2). Some tags occurred frequently: protein_structure (140), no-tag (134), and protein (114). By journal, tags were: docking (Proteins, 85), no-tag (JAMA, 20), and protein_structure (J Mol Biol, 52). No-tag (system assigned) indicated no tag assigned. Descriptors were more heavily reused than tags, for example: 'Models, Molecular' (550), Protein Conformation (363), and Humans (341). By journal, descriptors were: 'Models, Molecular' (Proteins, 252), 'Models, Molecular' (J Mol Biol, 235), and Humans (JAMA, 137). DISCUSSIONS AND CONCLUSIONS Comparison of tag and descriptor lists shows many of the same similarities and differences as the previous study [2]. Many user terms were related to the author and intermediary terms but not in the thesaurus (e.g. 'diet' and 'fat' used separately in the tag lists where they were linked as 'dietary fats' in the thesaurus). Terms such as 'human' and 'family-studies' show users tagging biology articles are interested in methodology and user groups associated with articles. This study has system design implications for accessing, indexing and searching document spaces. Users express frustration trying to narrow search results. Controlled vocabularies help narrow a search to a manageable size but can be expensive. User tagging could provide additional access points to traditional controlled vocabularies and the associative classifications necessary to tie documents and articles to time and task relationships among other novel items. REFERENCES [1] Tang H, Ng J.HK. 2006. Googling for a diagnosis -- use of Google as a diagnostic aid: internet based study. BMJ 333 (2 Dec), 1143-1145. [2] Kipp MEI. 2006. Complementary or discrete contexts in online indexing: A comparison of user, creator, and intermediary keywords. Canadian Journal of Information and Library Science (in press) http://dlist.sir.arizona.edu/1533/

Geospatial information technologies revolutionize the way we have traditionally approached navigation and browsing in information systems. Colorful graphics, statistical summaries, geospatial relationships of underlying collections make them attractive for text retrieval systems. This paper examines the nature of georeferenced information in academic library catalogs organized according to the Library of Congress Classification (LCC) with the goal of understanding their implications for geovisualization of library collections.

Artificial Intelligence Lab, Department of MIS, University of Arizona / Much of the work in business and academia is performed by groups of people. While significant advancement has been achieved in enhancing individual productivity by making use of information technology, little has been done to improve group productivity. Prior research suggests that we should know more about individual differences among group members as they respond to technology if we are to develop useful systems that can support group activities. We report results of a cognitive study in which researchers were observed performing three complex information entry and indexing tasks using an Integrated Collaborative Research System. The observations have revealed a taxonomy of knowledge and cognitive processes involved in the indexing and management of information in a research collaboration environment. A detailed comparison of knowledge elements and cognitive processes exhibited by senior researchers and junior researchers has been made in this article. Based on our empirical findings, we have developed a framework to explain the information management process during research collaboration. Directions for improving design of Integrated Collaborative Research Systems are also suggested.

Conference is sponsored by the Library of Congress Cataloging Directorate. / The decade of the 1990s saw the development of a proliferation of metadata element sets for resource description. This paper looks at a subset of these metadata schemes in more detail: the TEI header, EAD, Dublin Core, and VRA Core. It looks at why they developed as they did, major points of difference from traditional (AACR2/MARC) library cataloging, and what advantages they offer to their user communities. It also discusses challenges to implementers of these schemes and possible future developments. It goes on to identify some commonalties among these cases, and to attempt to generalize from these some lessons for developers of metadata element sets. It concludes by suggesting we also look carefully at emerging schemes being developed by publishers in support of electronic commerce and rights management, and think seriously about the implications of commodity metadata upon our traditional bibliographic apparatus.