Journal of the American Society for Information Science (JASIS) -- Table of Contents

Contributed by
Richard Hill
American Society for Information Science
Silver Spring, Maryland, USA
[email protected]

VOLUME 50, NUMBER 1 (JANUARY 1999)

CONTENTS

  • EDITORIAL
    The Coming of the Millenium and the Journal of the American Society for Information Science (JASIS)

    Donald H. Kraft
  • In This Issue
    Bert R. Boyce

    The three articles contributed to this special topic issue on youth issues are introduced by the editor of the issue. The other four articles in this issue are covered here.

  • Special Topic Issue: Youth Issues in Information Science
    Guest Editors: Mary K. Chelton and Nancy P. Thomas
  • Introduction: Why a Special Topic Issue on Youth Issues?
    Mary K. Chelton and Nancy P. Thomas
  • Utilization of Heroin Information by Adolescent Girls in Australia: A Cognitive Analysis
    Ross J. Todd
  • A Visit to the Information Mall: Web Searching Behavior of High School Students
    Raya Fidel, Rachel K. Davies, Mary H. Douglass, Jenny K. Holder, Carla J. Hopkins, Elisabeth J. Kushner, Bryan K. Miyagishima, and Christina D. Toney
  • Barriers to Adolescents' Information Seeking for Career Decision Making
    Heidi E. Julien

RESEARCH

  • A Re-Unification of Two Competing Models for Document Retrieval
    David Bodoff

    Standard vector space relevance feedback uses relevance information only to modify the particular query vector. If one uses both document and query adjustment, one seems to assume that query terms and document terms are both fixed and estimated at the same time. By assuming that neither are fixed and all must be estimated, Bodoff suggests a new combined model using multi-dimensional scaling techniques to minimize the differences between feedback data and distance between point pairs, while fixing enough data points to prevent arbitrary rotations, and thus overcoming the arbitrary positioning of the relatively optimized result. Alternatively the M.D.S. procedure may be run to optimize on relevance data without fixing points, and then rotated to minimize the distance to the original observed data points, or finally, both sources of data may be combined into a single function whose correct form may only be hypothesized. The thought is to accumulate relevance data off-line until enough is available to justify a new scaling run. A simulation which first corrects queries, then documents, has been run, and suggests considerable improvement. Data for a field study are being gathered.

  • Exploiting Parallelism in a Structural Scientific Discovery System to Improve Scalability
    Gehad M. Galal, Diane J. Cook, and Lawrence B. Holder

    Galal, Cook, and Holder look at the potential of using a computationally intensive knowledge discovery, data mining system (SUBDUE) with very large databases by way of parallel and distributed resources. With a functional approach an increase in processors not only reduces processing time but allows more substructures to be evaluated with a possible increase of compression. The time required to generate and evaluate initial structures prior to distribution to the processors imposes a limit on benefit. Dynamic partitioning showed little improvement, but static partitioning handles large databases with fewer resources than the sequential version and discovers equal or better substructures.

  • Inner-City Gatekeepers: An Exploratory Survey of Their Information Use Environment
    John Agada

    Twenty community organizers and organizational leaders were identified by community-based organizations as gatekeepers of information, in the sense that they could link community members with alternatives and solutions dependent to some extent on external cultures. Agada studied the Information Use Environment (IUE) of these gatekeepers by 1.5-hour recorded, structured interviews on demographics, problem situations, sources of unresolved problems, and preferred information sources.

    Gatekeepers have a higher level of education and income than the rest of the community but are otherwise little different. The categories of information needs used by Dervin are present but rank differently in priority with an apparent emphasis on coping skills. The gatekeepers believe the information they need exists but is often unavailable due to lack of awareness, access, or proficiency. Interpersonal communication seems to be a very important source with a preference for filtering through a member of the gatekeeper 's subculture.

  • Boolean Search: Current State and Perspectives
    Valery I. Frants, Jacob Shapiro, Isak Taksa, and Vladimir G.Voiskunskii

    Common objections to the Boolean retrieval model are reviewed by Frants et al. with a critical eye. If the use of operators is unfriendly to users, one might think that the several conversion algorithms from free text to Boolean statements which have been demonstrated might have been implemented. If ranked output is a serious problem for users, one might wonder why the several ranking methods in the literature for Boolean output have not been widely implemented. Methods exist for the automatic control of retrieved set size, which call into question the criticism that Boolean systems do not control output size. There is evidence that inverted file maintenance is more efficient than that of clustered files, and a number of algorithms exist for automatic feedback in Boolean systems.

    It is the published criticism of Boolean systems which has stimulated the development of improvements, but there have also been adverse effects of such criticism. Because a large percentage of the research community seems to believe that Boolean systems are in some way unsound, they receive little funding support, are difficult to use as a basis for publication, and are not an area encouraged for new research.

Click here to return to the D-Lib Magazine clips column.