Information Retrieval List Digest 104 (March 16, 1992) URL = http://hegel.lib.ncsu.edu/stacks/serials/irld/irld-104 IRLIST Digest March 16, 1992 Volume IX, Number 8 Issue 104 ********************************************************** I. NOTICES A. Meeting Announcements/Calls for Papers 1. Data Administration Symposium, May 12-13, 1992, Gaithersburg, MD 2. Information and Knowledge Management, November 8-11, 1992, Baltimore, MD II. QUERIES B. Requests for Information 1. Information Access 2. Dictionary Design and Maintenance Request IV. PROJECT WORK C. Abstractsphies 1. IR-Related Dissertation Abstracts ********************************************************** I. NOTICES I.A.1. Fr: Barry L. Newton Re: DAMA Symposium Meeting Notice ANNOUNCING THE FIFTH ANNUAL DAMA SYMPOSIUM DATA ADMINISTRATION - A VALUE ADDED SERVICE MAY 12 - 13, 1992 NIST, GAITHERSBURG, MD The National Capital Region Chapter of the Data Administration Management Association proudly announces its fifth annual Symposium. This year, we emphasize the practices, technologies, activities, initiatives, and ideas that deliver clearly visible value to users, or "customers" of Data Administration - and could make a difference in your organization and build support for your Data Administration program. To address this theme, the Symposium features: Three nationally-known Keynote speakers: Barbara von Halle, Charlie Bachman, and Dan Appleton; Other speakers relating real life experiences of Data Administration, including: Cindy Walker of Software Solutions; Andrea Norris of NASA; Jerry Winkler of ASYSA, Chair of the ANSI IRDS Standard Committee; and Cathy Hirsh of AMS; Panel discussions on topics including CASE, the System Life Cycle, and IRDS Implementation; And Workshops on Data Standardization, Automated Tools, Security, and Repository Management. The registration fees (by mail by April 21), are $200.00 for current DAMA members and $250.00 for non-DAMA members. This includes refreshments, luncheons, presentation materials, and 1992-1993 DAMA membership. Government training forms are accepted. Send fee to: DAMA-NCR Symposium Registration P.O. Box 598 Columbia MD 21045-0598 For more information or a brochure call: Julie Benoit 703-521-5300 ********** I.A.2. Fr: Timothy Finin Re: CFP: CIKM-92 -- Conf. on Information and Knowledge Management C A L L F O R P A P E R S First International Conference on INFORMATION AND KNOWLEDGE MANAGEMENT November 8 - 11, 1992 Radisson Hotel, Baltimore, Maryland, USA Sponsored by ISMM in cooperation with AAAI, SIGART, SIGIR, IEEE The conference provides an international forum for presentation and discussion of research on information and knowledge management, as well as recent advances on data and knowledge bases. Authors are invited to submit papers and proposals for tutorials concerned with theory or practice or both. The focus of the conference includes but is not limited to, the following: application of knowledge representation techniques to semantic data modeling; development and management of heterogeneous knowledge bases; automatic acquistion of data and knowledge bases especially from raw text; object-oriented DBMS; optimization techniques; transaction management; high performance OLTP systems; security techniques; performance evaluation; hypermedia; unconventional applications; parallel database systems; physical and logical database design; data and knowledge sharing; interchange and interoperability; cooperation in heterogenous systems; domain modeling and ontology-building; knowledge discovery in databases; information storage and retrieval and interface technology. INSTRUCTIONS TO AUTHORS: All submissions must be accompanied by a cover letter containing a list of all authors, their affiliations, telephone numbers, email addresses, and fax telephone numbers. Papers should be at most most 20 double-spaced pages and must include an abstract of 100-150 with five keywords. All submissions will be reviewed and will be judged with respect to quality and relevance. The author of the best student paper will receive an award for his/her submission. To be eligible, the student must be the first author and primary contributor to the paper. The cover letter must identify the paper as a candidate for this competition. Submit five copies of each paper or tutorial proposal to the program chair: Dr. Yelena Yesha Computer Science Department University of Maryland Baltimore County Telephone: +1 410-455-3000 5401 Wilkens Ave Fax: +1 410 455-3969 Baltimore, Maryland, 21228-5398 Email: cikm@cs.umbc.edu IMPORTANT DATES: Deadline for paper and/or tutorial submission: May 31, 1992 Notification of acceptance: July 15, 1992 Camera ready papers due: August 10, 1992 *************************************************************************** ** For an automatic reply containing more information about CIKM-92 send ** ** mail to cikm-info@cs.umbc.edu. Send queries to cikm@cs.umbc.edu. ** *************************************************************************** CONFERENCE CO-CHAIRS Tim Finin (USA) A. F. Norcio (USA) PROGRAM CHAIR Yelena Yesha ORGANIZING COMMITTEE Bruce Blum Keith Humenik John Meisenbacher E. K. Park PROGRAM COMMITTEE Raphael Alonso (USA) Ed Fox (USA) Mark Fox (Canada) George Gardarin (France) Larry Kerschberg (USA) Mike Liu (USA) Peter Matthews (USA) James Mayfield (USA) Don McKay (USA) Daniel Menasce (USA) Bob Neches (USA) Charles Nicholas (USA) Chris Overton (USA) Mike Papazouglou (AUS) Yun Peng (USA) Larry Reeker (USA) Karen Ryan (USA) Ralph Semmel (USA) Stu Shapiro (USA) Amit Sheth (USA) Deepinder Sidhu (USA) Mukesh Singhal (USA) Il-Yeol Song (USA) John Sowa (USA) Brooke Stephens (USA) Alex Thomasian (USA) Patrick Valduriez (France) Lawrence A. Welsch (USA) ********************************************************** II. QUERIES II.B.1. Fr: Marc Sullivan Re: Information Access My speciality is information retrieval and information filtering. Currently I'm working with an object-oriented database person to implement advanced retrieval techniques in an OODB. I would like to discuss general problems of people getting access to information as well as new algorithms, etc. ********** II.B.2. Fr: marwk@levels.unisa.edu.au Re: Dictionary design and maintentance request There are various methods which can be used to store a dictionary of 80,000 words. Speed of access and size of file are 2 criteria. Some dictionaries can be maintained by the user, while other supply a supplementary dictionary. Hashing, encoding, ... can be used. Please inform me of (1) bibiliographic information relating to this area (2) example source code for this (3) information on the methods of various software packages, e.g. WORD PERFECT. Thank you. Raymond University of South Australia | Act in haste and repent at leasure P.O. Box 1 | Code too soon and debug forever Ingle Farm | Knobs, knobs everywhere, South Australia | just vary a knob to think! ********************************************************** IV. PROJECT WORK IV.C.1. Fr: Susanne M. Humphrey Re: Selected IR-Related Dissertation Abstracts The following are citations selected by title and abstract as being related to Information Retrieval (IR), resulting from a computer search, using BRS Information Technologies, of the Dissertation Abstracts Online database produced by University Microfilms International (UMI). Included are UMI order number, title, author, degree, year, institution; number of pages, one or more Dissertation Abstracts International (DAI) subject descriptors chosen by the author, and abstract. Unless otherwise specified, paper or microform copies of dissertations may be ordered from University Microfilms International, Dissertation Copies, Post Office Box 1764, Ann Arbor, MI 48106; telephone for U.S. (except Michigan, Hawaii, Alaska): 1-800-521-3042, for Canada: 1-800-268-6090. Price lists and other ordering and shipping information are in the introduction to the published DAI. An alternate source for copies is sometimes provided. Dissertation titles and abstracts contained here are published with permission of University Microfilms International, publishers of Dissertation Abstracts International (copyright by University Microfilms International), and may not be reproduced without their prior permission. AN University Microfilms Order Number ADGDX-92394. AU BARNARD, JOHN MORDAUNT. TI THE COMPUTER STORAGE, RETRIEVAL AND SEARCHING OF GENERIC STRUCTURES IN CHEMICAL PATENTS: THE MACHINE-READABLE REPRESENTATION OF GENERIC STRUCTURES. IN University of Sheffield (United Kingdom) Ph.D. 1982, 400 pages. DE Information Science. Chemistry, General. Computer Science. AB Available from UMI in association with The British Library. The nature of the generic chemical structures found in patents is described, with a discussion of the types of statement commonly found in them. The available representations for such structures are reviewed, with particular note being given to the suitability of the representation for searching files of such structures. Requirements for the unambiguous representation of generic structures in an "ideal" storage and retrieval system are discussed. The basic principles of the theory of formal languages are reviewed, with particular consideration being given to parsing methods for context-free languages. The Grammar and parsing of computer programming languages, as an example of artificial formal languages, is discussed. Applications of formal language theory to chemistry and information work are briefly reviewed. GENSAL, a formal language for the unambiguous description of generic structures from patents, is presented. It is designed to be intelligible to a chemist or patent agent, yet sufficiently formalised to be amenable to computer analysis. Detailed description is given of the facilities it provides for generic structure representation, and there is discussion of its limitations and the principles behind its design. A connection-table-based internal representation for generic structures, called an ECTR (Extended Connection Table Representation) is presented. It is designed to represent generic structures unambiguously, and to be generated automatically from structures encoded in GENSAL. It is compared to other proposed representations, and its implementation using data types of the programming language Pascal described. An interpreter program which generates an ECTR from structures encoded in a subset of the GENSAL language is presented. The principles of its operation are described. Possible applications of GENSAL outside the area of patent documentation are discussed, and suggestions made for further work on the development of a generic structure storage and retrieval system based on GENSAL and ECTRs. AN University Microfilms Order Number ADG91-16241. AU LIU, MENGXIONG. TI THE CITATION PROCESS IN SCIENTIFIC COMMUNICATION: AN ANALYSIS OF CITER MOTIVATION AND CITATION CHARACTERISTICS OF CHINESE PHYSICISTS. IN The University of Michigan Ph.D. 1990, 224 pages. DE Information Science. Library Science. AB A crucial aspect in the creation and dissemination of information is the use of citation--references made in published books and articles to the writings of other authors. Numerous studies have sought to understand the relationships between the practice of citation and those documents cited in scientific literature. Methodologies of those studies, however, depended heavily on inference rather than empirical evidence. Moreover, they took two distinct approaches. The normative school claimed that the citation process adhered to standards, while the opposing side contended that other factors were involved. The goal of this study was to obtain insights into the citation process in scientific communication focusing on a barely touched problem--citation motivation. The hypothesis was tested that the citation process is not conducted with adherence to standards, rather it is associated with extra-scientific effects, and that citing motivators are most likely to contribute to the underlying norms of the citation process. Different from most inferential methodologies, the findings of this research were derived from directly questioning the Chinese physicist authors who cited other authors' works in their published papers. Indicators of citing characteristics and motivations were identified by analyzing the data collected from the questionnaires. As a result, internal motivators and external factors were determined in terms of their relationships to the two citation indicators: citation output and citation essentiality. Findings revealed that an author's citation output was positively related to his use of an institutional library, while citation essentiality was related to an author's personal assessment of the value of the citation to a particular scientific domain. Moreover, a single citation could serve a number of functions in both personal motivation and in social or psychological factors. The results of this study partially supported its hypothesis that the citation process does not necessarily conform to standards nor does it operate under rational evaluation. The study concluded that an author's citing process is unique, personal and complex. It suggested that further investigations are needed to articulate the nature and norms of this more-private-than-public process. The significance of this study lies in its exploratory approach to the problem of citing behavior and citer motivation via a non-traditional methodology--direct questioning, and in its achievement of some new insights into the citation process. AN University Microfilms Order Number ADG91-16167. AU DIMITROFF, ALEXANDRA. TI MENTAL MODELS AND ERROR BEHAVIOR IN AN INTERACTIVE BIBLIOGRAPHIC RETRIEVAL SYSTEM. IN The University of Michigan Ph.D. 1990, 206 pages. DE Library Science. Information Science. AB The purpose of this study was to examine the relationship between the end-user searcher's mental model of an interactive bibliographic retrieval system and success in searching the system as reflected in error behavior. This study examined the level of mental model completeness as it related to error behavior and search success. Interviews were conducted to determine the level of completeness of each subject's mental model of an interactive bibliographic retrieval system. Subjects were categorized in terms of complete, good, incomplete and poor mental models. Subjects searched an interactive bibliographic retrieval system for a predetermined set of items. Analysis of variance and means separation tests were used in data analysis. Results supported the hypothesis that there is a relationship between the completeness of the end-user researcher's mental model and error behavior as well as total number of items found. Subjects with more complete mental models made significantly fewer errors and found significantly more items. However, the majority of subjects had incomplete or poor mental models, suggesting that the development of more complete mental models is an issue that must be addressed, either through training or system supplied assistance. The most frequent types of errors made were errors consisting of the use of too broad a search strategy followed by subject searching errors. Both of these error types could be avoided by system-provided cues, cues not given by the system used in this study. Statistically significant relationships existed between errors made and the following: computer tasks, academic field, frequency of computer use, and frequency of system use. Search success was found to be significantly related to computer hardware, computer tasks, database design experience, academic field, frequency of computer use, and frequency of system use. Subjects who had learned to use personal computers without formal training found significantly more items than those who had taken a formal course. Conclusions and implications for interactive bibliographic retrieval system design and training are discussed. Recommendations for future research are proposed. AN University Microfilms Order Number ADG91-17444. AU WOOD, JO ANNA. TI A MODEL OF STRATEGIES IN USING EXTERNAL INFORMATION. IN Southern Illinois University at Carbondale Ph.D. 1990, 192 pages. DE Psychology, Experimental. Information Science. AB People often have problems searching external sources for information. Failure to find the needed information can be devastating in some circumstances (e.g., in military operations, or medicine). With the growth of computerized information stores, the problem is compounded by a searcher's difficulties using search tools. This paper presents and tests a model of how people search external information, in the form of a computerized database, in a sentence verification task. The model was based on Reder's (1987) model of memory search, and Kolodner's computer model of executive search strategies. The model was tested on sixteen subjects and latency estimates for each of the model's components were derived for each subject. After the initial model failed to meet acceptability criteria, a revised model was tested. The revised model yielded satisfactory results for fifteen of the subjects. The revised model also compared favorably to simpler, more parsimonious models. Results indicate that the model provides a reasonable description of the decisions that are made during the search process. Decision parameter estimates are compared to Reder's results. Discussion focuses on the theoretical and practical implications of the model and some extensions to it. ********************************************************** IRLIST Digest is distributed from the University of California, Division of Library Automation, 300 Lakeside Drive, Oakland, CA. 94612-3550. Send subscription requests to: LISTSERV@UCCVMA.BITNET Send submissions to IRLIST to: IR-L@UCCVMA.BITNET Editorial Staff: Clifford Lynch lynch@postgres.berkeley.edu or calur@uccmvsa.bitnet Nancy Gusack ncgur@uccmvsa.bitnet Mary Engle engle@cmsa.berkeley.edu or meeur@uccmvsa.bitnet The IRLIST Archives will be set up for anonymous FTP, and the address will be announced in future issues. To access back issues presently, send the message INDEX IR-L to LISTSERV@UCCVMA.BITNET. To get a specific issue listed in the Index, send the message GET IR-L LOG ***, where *** is the month and day on which the issue was mailed, to LISTSERV@UCCVMA.BITNET. These files are not to be sold or used for commercial purposes. Contact Nancy Gusack or Mary Engle for more information on IRLIST. The opinions expressed in IRLIST do not represent those of the editors or the University of California. Authors assume full responsibility for the contents of their submissions to IRLIST. Received: by stubbs.ucop.edu (5.57/1.34) id AA21337; Thu, 18 Feb 93 14:43:34 -0800 Message-Id: <9302182243.AA21337@stubbs.ucop.edu> Received: from UCCVMA.UCOP.EDU by uccvma.ucop.edu (IBM VM SMTP V2R2) with BSMTP id 8852; Thu, 18 Feb 93 14:49:19 PDT Received: from UCCMVSA.BITNET (NJE origin NCG$UR@UCCMVSA) by UCCVMA.UCOP.EDU (LMail V1.1c/1.7e) with BSMTP id 2688; Thu, 18 Feb 1993 14:49:18 -0700 Received: by UCCMVSA.BITNET Thu, 18 Feb 93 14:48:55 PST Date: Thu, 18 Feb 93 14:48:55 PST From: "Nancy Gusack" To: NCG@stubbs Status: O IRLIST Digest March 24, 1992 Volume IX, Number 9 Issue 105 ********************************************************** I. NOTICES A. Meeting Announcements/Calls for Papers 1. European Conference on Hypertext, '92, November 30 - December 4, 1992, Milano, Italy B. Publication Announcements 1. _EJournal_ 2. Online Information Hunting IV. PROJECT WORK C. Abstracts 1. IR-Related Dissertation Abstracts ********************************************************** I. NOTICES I.A.1. Fr: Robert.Akscyn@zog.cs.cmu.edu Re: CAll for ECHT '92 Preliminary Call for Papers European Conference on Hypertext (ECHT'92) Fourth ACM Conference on Hypertext 30 November - 4 December 1992 Milano, Italy SCOPE: ECHT'92 is the second in a series of European conferences on Hypertext and Hypermedia in alternation with the U.S. based HYPERTEXT'xx conferences under coordination of ACM SIGLINK. ECHT'92 is a major event where researchers, developers and users can meet around the theme of Hypertext and Hypermedia. The broad applicability of Hypertext and Hypermedia as a primary technology in many domains and its efficiency as an information integrator has lead to increase interest from industry as well as recognition from academia over the last five years. By its very nature, Hypertext and Hypermedia are at the intersection of many fields within computer science and also cognitive science, education, communication, and are relevant for many application domains. Therefore, ECHT'92 will be of interest for a broad spectrum of professionals ranging from theoreticians to system and application developers, from researchers to authors and end-users. The conference will include prominent guest speakers, presentations of refereed papers, panel sessions, technical briefings sessions, poster and video presentations, as well as demonstrations of experimental research prototypes and commercial products. The conference will feature two days tutorials of introductory and advanced course level. There will be opportunities for informal meetings of special interest groups. TOPICS: We are inviting you to participate in ECHT'92 and to submit original and innovative contributions such as papers, panels, tutorials, demos, videos and posters. Submissions in any area concerned with Hypertext and Hypermedia research and development are encouraged. A non exhaustive list of topics of interest is suggested: Hypertext and Hypermedia - applications - modelling and design - development methodologies and tools - responsive interfaces - evaluation - Systems software technologies - authoring Hypertext-Hypermedia in connection with: - Operating Systems - Data Base Management Systems - Object-Oriented Systems and languages - Knowledge Systems - Information retrieval - Software engineering - Cooperative work - Fiction CONFERENCE CO-CHAIRs Paolo Paolini Politecnico di Milano, Dipartimento di Elettronica, E-mail: PAOLINI@IPMEL1.POLIMI.IT Telephone: (39) 2-23993520 Fax: (39) 2-23993411 Gianni Degli Antoni Universita di Milano (Italy) PROGRAM COMMITEE CO-CHAIRS Jocelyne & Marc Nanard LIRMM, University of Montpellier, E-mail: nanard@crim.fr Telephone: (33) 67 63 04 60 Fax: (33) 67 54 73 76 TUTORIALS CHAIR Franca Garzotto Politecnico di Milano PANELS CHAIR Norbert Streitz GMD-IPSI VIDEOS and DEMOS CHAIR Paul Kahn IRIS - Brown Dynamic Diagrams Inc. TECHNICAL BRIEFINGS CHAIR Norman Meyrowitz GO Corporation ECHT '92 PROGRAM COMMITEE (To be extended) Robert Akscyn Knowledge System (USA) Patricia Baird Scottish Daily Record (UK) Mark Bernstein Eastgate Systems Inc.(USA) Peter Brown Univ. of Kent (UK) Tat Seng Chua Ntl. Univ. of Singapore (Singapore) Ralf Cordes Telenorma Bosch Telecom (FRG) Franca Garzotto Politecnico di Milano (I) Nuno Guimaraes INESC(P) Frank Halasz Xerox Parc (USA) Hiroshi Ishii NTT (Japan) Paul Kahn IRIS, Brown Univ. (USA) John Leggett Texas A&M Univ. (USA) Dario Lucarella ENEL (I) Norman Meyrowitz GO Corporation (USA) Jakob Nielsen Bellcore (USA) Tim Oren Apple (USA) Antoine Rizk Euroclid (F) Daniel Schwabe PUC (Brazil) Norbert Streitz GMD-IPSI (FRG) Frank Tompa Univ. of Waterloo (CAN) Randall H. Trigg Aarhus Univ (DK) A. M. Vercoustre INRIA (F) Janet Walker DEC Cambridge Res. Lab. (USA) N. Yankelovich SUN (USA) ECHT'92 CONFERENCE SECRETARIAT Enza Caputo Politecnico di Milano, Dipartimento di Elettronica, Piazza Leonardo da Vinci 32 20133 Milano (Italia) E-mail: Caputo@IPMEL1.POLIMI.IT Telephone: (39) 2-23993405 Fax: (39) 2-23993411 INSTRUCTIONS FOR SUBMISSION: (to Conference Secretariat) Full papers (<6000 words) should be submitted in five paper copies. A separate cover page must contain title of paper, name(s), affiliation and complete mailing address (incl. phone, telefax, e-mail) of the authors together with an abstract (about 200 words) and (from three to five) keywords. Proposals for panels should contain topic of the panel, its content (<1000 words), names and affiliation of proposer and contributors, a brief characterization of their background and experiences relevant to the panel. Proposals for videos, demos and posters presentations should be submitted in the form of an extended abstract, describing the content and also the content of the course and its format (1000-2000 words), should identify the target audience, the level of expertise required, and the lenght (1 or 2 half days). Qualification and profile of the tutor(s) should also be included. DEADLINES May 4th 1992: Submission of full papers, panels, videos, demos, posters, and tutorials should be received by the conference secretariat. Please, send also an Email version of the abstract with title, name, address and affiliation, to the conference secretariat as soon as possible, and in any case before the deadline. July 6th 1992: Notification of acceptance. Sept. 1st 1992: Final copy of papers received by the conference secretariat. ********** I.B.1. Fr: EJOURNAL@ALBNYVMS.BITNET Re: _EJournal_ Announcing an Electronic Journal March 1992 _EJournal_ is a peer-reviewed, all-electronic, network distributed, serial publication. We are particularly interested in theory and practice surrounding the creation, transmission, storage, interpretation, alteration and replication of electronic "text," broadly defined. We are also interested in the social, psychological, literary, economic, pedagogical, philosophical and other ramifications of computer-mediated networks. Our review process is anonymous, all-electronic, and consensual. We prefer brief, authentic, lively essays to exhaustive technical reports. Single-essay issues appear as often as submissions are affirmatively reviewed; there were four (free) distributions to subscribers in 1991. There are two dozen consulting editors, in several disciplines, who review submissions. Members of _EJournal_'s advisory board are: Stevan Harnad, Princeton University Dick Lanham, University of California at Los Angeles Ann Okerson, Association of Research Libraries Joe Raben, City University of New York Bob Scholes, Brown University Harry Whitaker, University of Quebec at Montreal To subscribe to _EJournal_, send a mail message to listserv@albnyvm1.bitnet containing as its only line the command: subscribe ejrnl your_first_name your_last_name Information about getting back issues will accompany the "Welcome" message sent to people who subscribe. Please send submissions for editorial consideration to our "office" at: ejournal@albnyvms.bitnet Ted Jennings, Editor, Department of English, University at Albany/SUNY ********** I.B.2. Fr: Nahum (N.) Goldmann Re: OIH ONLINE INFORMATION HUNTING, by Nahum Goldmann This book (McGraw-Hill/Windcrest, ISBN 0-8306-3945-4, tel. 1-800-233-1128, price Paper $17.95, Hard $29.95) is amongst a limited number of publications on computerized information retrieval specifically addressed to the end-user. The book describes what kind of information is available online, how to search for it, and how to use it to your professional advantage. The book outlines the SUBJECT EXPERT SEARCHING TECHNIQUE - a new methodology for online information gathering. This efficient research method is especially tailored for the end-users of information who are working in rapidly developing scientific areas, and can be used as a means of professional survival in rapidly changing scientific and professional areas. It has been successfully used by several R&D organizations. With this technique, occasional users such as scientists and students can now master the use of online databases for themselves. "... Although intended for end-users, the book could serve as... an introductory online search course for librarians." - Dr. Roger K. Summit, Dialog "... I recommend that practicing physicians, researchers, and students start with this very useful book." - Can. Med. Ass. J "... A definite basis for an undergraduate course in '(Intelligent) Information Retrieval.'" - Dr. Alex Meystel, Drexel U. ********************************************************** IV. PROJECT WORK IV.C.1. Fr: Susanne M. Humphrey Selected IR-Related Dissertation Abstracts The following are citations selected by title and abstract as being related to Information Retrieval (IR), resulting from a computer search, using BRS Information Technologies, of the Dissertation Abstracts Online database produced by University Microfilms International (UMI). Included are UMI order number, title, author, degree, year, institution; number of pages, one or more Dissertation Abstracts International (DAI) subject descriptors chosen by the author, and abstract. Unless otherwise specified, paper or microform copies of dissertations may be ordered from University Microfilms International, Dissertation Copies, Post Office Box 1764, Ann Arbor, MI 48106; telephone for U.S. (except Michigan, Hawaii, Alaska): 1-800-521-3042, for Canada: 1-800-268-6090. Price lists and other ordering and shipping information are in the introduction to the published DAI. An alternate source for copies is sometimes provided. Dissertation titles and abstracts contained here are published with permission of University Microfilms International, publishers of Dissertation Abstracts International (copyright by University Microfilms International), and may not be reproduced without their prior permission. AN University Microfilms Order Number ADG91-19523. AU BHUYAN, JAYANARAYAN. TI CLUSTER-BASED ADAPTIVE INFORMATION RETRIEVAL. IN University of Southwestern Louisiana Ph.D. 1990, 194 pages. SO DAI V52(02), SecB, pp925. DE Computer Science. AB The objective of this dissertation is to design and evaluate a complete Information Retrieval system in which the feedback from a number of users of the system, about its performance (global feedback), is stored in the form of clusters called user-oriented clusters. These clusters are utilized in answering a query in conjunction with feedback obtained about the system's performance after a partial retrieval for that query (local feedback). The classification schemes are developed in 2 stages. The first stage involves the accumulation of relevance judgement provided by users vis-a-vis past query instances. The second stage consists of the identification of clusters. In one of these schemes, the feedback is accumulated on a straight line and the cluster identification problem is termed as Boundary Selection Problem (BSP). Several solutions to this problem, the first with the use of a Greedy approach, the second with the use of Dynamic Programming, and the third with the combination of Branch and Bound algorithm with Dynamic Programming, are developed. In the other scheme, the feedback is accumulated on an undirected graph and an optimization function that measures the effectiveness of the formed clusters is developed. This optimization problem is found to be NP-hard and three heuristic approaches, which are based respectively on a Greedy approach, Simulated Annealing technique, Genetic algorithm and Simulated Evolution, are developed. Once the clusters are formed, we develop two methods for searching the clusters relevant to a user's query. In one of the methods, which we call Retrieval without cluster characterization, a probabilistic model in which we utilize the information in the feedback accumulation stage to determine the probability of relevance of a query to a cluster is developed. In another method we take the description of documents into account in order to describe the clusters, and use traditional vector space model for retrieval purposes. Our experiments with clustering methods show that the clusters generated by our method outperform existing methods for user-oriented clustering. The performance of our overall system is compared to that of conventional vector space model and is found to be superior in almost all cases. AN University Microfilms Order Number ADG91-18230. AU GORMAN, KEVIN JOSEPH. TI AN OBJECT-ORIENTED SEMANTIC DATA MODEL. IN Texas A&M University Ph.D. 1990, 305 pages. SO DAI V52(02), SecB, pp930. DE Computer Science. Business Administration, Management. Information Science. AB This research centers on the design and validation of The Object-Oriented Semantic Data Model (OOSDM). OOSDM integrates features and concepts from semantic data models, object-oriented programming languages, and object-oriented data models; resulting in a data model supporting: (1) An abstract data type (APT) facility; (2) Transaction specification; (3) Enhanced logical and procedural derivation; (4) Explicit constraint specification; (5) Operational semantics for manipulating 'complex objects'. OOSDM includes additional components lacking in existing object-oriented data models: a query language and an associated diagramming technique. The Object-Oriented Semantic Query Language (OOSQL) extends relational calculus based query languages with: logical access path independence, and the use of APT operations and generalization hierarchies in queries. OOSQL contributes a formally defined, high-level, calculus based query language. Current object-oriented data models possess no such object-oriented calculus (OOC) counterpart to relational calculus. Object-Oriented Entity-Relationship Diagrams (OOERDs) extend the popular Entity-Relationship approach for modeling object structure. OOERDs model object dynamics for all operations: complex object, APT, and query. OOERDs furnish an additional database design aid, providing a graphical abstraction of both structural and behavioral application characteristics. In validating OOSDM, three separate formal mappings furnish a theoretical basis: (1) A structural mapping of OOSDM schemes to relational model schemes; (2) An operational mapping of OOC to relational calculus; (3) A mapping of access path independent OOC to access path dependent OOC. For all three mappings, justification is provided for claims of completeness, consistency, and correctness. The structural and operational mappings contribute the only known formal mappings of a semantic model to the relational model. OOSDM's access path mapping is an original approach to achieving logical access path independence. The mapping employs a semantic model graph and a simple breadth-first search algorithm to define the access path(s). Advantages of the approach include: (1) Applicability to both object-oriented models and to the relational model through a semantic interface; (2) Support for user-selected query interpretation; (3) Algorithm simplicity; (4) Database retrieval using only lossless join decompositions; (5) No additional required design concepts. AN University Microfilms Order Number ADG91-20689. AU LU, WEN YU. TI STRUCTURAL ANALYSIS AND ITS APPLICATION TO RECURSIVE QUERY PROCESSING IN DATABASES. IN The Ohio State University Ph.D. 1991, 147 pages. SO DAI V52(02), SecB, pp935. DE Computer Science. AB Efficient logic query processing has been an important issue in extending the traditional database systems into sophisticated database and knowledge base systems for the last decade. Single linear recursion has been identified as THE most common class of recursion in applications. The bottleneck to achieving efficiency is the repetitive, expensive, redundant operations in accessing the underlying databases in query processing. Significant efforts and many good results have been reported in the literature, yet a thorough understanding of linear recursion is yet to be seen. We conduct a detailed analysis on the structure and the expansions of a class of single linear recursion aiming at exploring the useful properties that can help reducing the processing redundancy. Our analysis proceeds as follows: (1) Analysis: A graph model of a linear recursion is adopted and the linear recursions are classified into four classes according to their configuration complexity. Then the two simpler classes are studied and characterized, with the conclusion used in deriving the properties and characterization of the two complex classes. Our characterization about different classes of recursions is categorized as bounded recursion, simple chain recursion and complex chain recursion. (2) Query processing: Based on the analysis, various ways of query optimization are discussed, with the emphasis on the realization of optimization principles: processing common subexpressions, performing selections and projections instead of joins. Then two processing algorithms are proposed for two important subclasses, namely, single linear recursion and multiple linear recursion, for selection queries. The algorithms are proved to be optimal in terms of the number of accesses to the underlying database. (3) Design of a query processor: The architecture of a query processor is proposed as a result of this study. The processor adopts the popular compilation approach in query processing and features a sophisticated analysis and compilation component that generates generic query processing plan to be instantiated by any related incoming queries. Rules are classified and characterized according to the study so that efficient processing plans can be generated. It also has the advantages of modularity and extensibility to make it easy for implementation. ********************************************************** IRLIST Digest is distributed from the University of California, Division of Library Automation, 300 Lakeside Drive, Oakland, CA. 94612-3550. Send subscription requests to: LISTSERV@UCCVMA.BITNET Send submissions to IRLIST to: IR-L@UCCVMA.BITNET Editorial Staff: Clifford Lynch lynch@postgres.berkeley.edu or calur@uccmvsa.bitnet Nancy Gusack ncgur@uccmvsa.bitnet Mary Engle engle@cmsa.berkeley.edu or meeur@uccmvsa.bitnet The IRLIST Archives will be set up for anonymous FTP, and the address will be announced in future issues. To access back issues presently, send the message INDEX IR-L to LISTSERV@UCCVMA.BITNET. To get a specific issue listed in the Index, send the message GET IR-L LOG ***, where *** is the month and day on which the issue was mailed, to LISTSERV@UCCVMA.BITNET. These files are not to be sold or used for commercial purposes. Contact Nancy Gusack or Mary Engle for more information on IRLIST. The opinions expressed in IRLIST do not represent those of the editors or the University of California. Authors assume full responsibility for the contents of their submissions to IRLIST.