IZ Research Reports

Former IZ research reports feature research results from the field of information sciences. They include doctoral dissertations or final reports on major projects. The reports were published by the former IZ publishing house.

Rough Set Theory based Automatic Text Categorization and the Handling of Semantic Heterogeneity

Xueying Zhang
Rough Set Theory based Automatic Text Categorization and the Handling of Semantic Heterogeneity

Social Science Information Centre, Bonn, Germany,
Nanjing University of Science and Technology, P.R. China and
Nanjing Normal University, P.R. China

Bonn: Informationszentrum Sozialwissenschaften 2006
Research reports; Volume 8
ISSN 1431-5114, ISBN 3-8206-0149-X

With the growing amount of information available in digital form and from a large number of sources, supporting technologies for automatically organizing this information are in high demand. Automatic text categorization, the automatic assignment of predefined categories to text documents with unknown content, therefore is currently the most popular text processing technique. Machine learning approaches for automatic text categorization, of which rough set theory (RST) is one which gains more and more attention, provide an alternative to traditional statistical approaches and those from knowledge engineering.
This report presents latest results in research on automatic text categorization focussing on language-independent text representation, the application of a RST model combining information retrieval theories and basic rough set theory for the treatment of semantic heterogeneity in information systems, and enhancing automatic text classification by rough set theory based text categorization, dynamic category extension and synonym detection.
The research report was accomplished within the framework of the Sandwich-Programme of the DAAD (German Academic Exchange Service) and out of cooperation between the Social Science Information Centre in Bonn and the Nanjing University of Sciences and Technology/ Nanjing Normal University, P.R. China. It is the English version of the Chinese thesis.

Visualisierung im Document Retrieval

Maximilian Eibl
Visualisierung im Document Retrieval
Theoretische und praktische Zusammenführung von Softwareergonomie und Graphik Design

2. Edition

Bonn: Informationszentrum Sozialwissenschaften 2003
(Research reports; Volume 7)
ISSN 1431-5114; ISBN 3-8206-0144-9

In der Frage nach der Ästhetik in der Ergonomie hat sich in den letzten Jahren ein Wandel vollzogen. Noch in den 1990er Jahren missachteten Ergonomen ästhetische Faktoren in der Gestaltung von Benutzungsschnittstellen weitgehend. Inzwischen hat sich allerdings das gegenseitige Interesse von Ergonomen und Designern verstärkt, so dass sich Herausgeber und Autor entschlossen, mit dem vorliegenden Band eine erweiterte und aktualisierte Ausgabe der vergriffenen ersten Auflage vorzustellen. Einige Teile der Erstauflage sind gestrafft, andere aktualisiert. Ferner wurden Teile hinzugefügt, die den aktuellen Stand der Thematik in Theorie und Praxis widerspiegeln. Der nun vorliegende Forschungsbericht enthält über 50 neue Textseiten, 39 neue Abbildungen und fast 90 neu eingearbeitete Referenzen.

Das Buch beschreibt detalliert die Visualisierung DEViD (Graphik Design und Softwareergonomie integrierende Visualisierung für Document Retrieval-Systeme) als Beispiel für eine Kooperation zwischen Ergonomie und Design. Neben weiteren praktischen Beispielen werden auch einige theoretische Überlegungen zur Integration von Design und Ergonomie vorgenommen.

Die beigelegte CD-ROM enthält

  • die in der Druckausgabe aus technischen Gründen schwarzweiß gehaltenen Abbildungen in Farbe
  • eine englische und deutsche Animation, die den Aufbau und die Funktionsweise von DEViD vorstellt
  • eine knapp 50seitige englischsprachige Zusammenfassung sowie
  • die Literaturlisten beider Auflagen.

Objektorientierte Dynamische Benutzungsoberflächen - ODIN

Maximilian Stempfhuber
Objektorientierte Dynamische Benutzungsoberflächen - ODIN 
Behandlung semantischer und struktureller Heterogenität in Informationssystemen mit den Mitteln der Softwareergonomie

Bonn: Informationszentrum Sozialwissenschaften 2003
(Research reports; Volume 6)
ISSN 1431-5114; ISBN 3-8206-0142-2

Integrierten Informationssystemen kommt bei der Informationsversorgung von Wissenschaft und Praxis eine immer stärkere Bedeutung zu. Die aktuell stattfindende Integration der vom Bundesministerium für Bildung und Forschung und Deutscher Forschungsgemeinschaft geförderten Informationsverbünde und Virtuellen Fachbibliotheken in das deutsche Wissenschaftsportal Vascoda zeigt dies deutlich. Dabei stellt die Unterschiedlichkeit der verfügbaren Datenbestände bezüglich ihrer formalen Struktur und inhaltlichen Erschließung die Entwickler von Informationssystemen vor neue Herausforderungen.

Der vorliegende Forschungsbericht beschreibt ODIN - ein Modell für die Gestaltung von grafischen Benutzungsoberflächen, das ein abgestimmtes Bündel von Konzepten und Oberflächenelementen zur Behandlung von Heterogenität darstellt.
ODIN führt zum einen eine neue Syntax für adaptive und adaptierbare Benutzungsoberflächen ein, die dem Nutzer sowohl auf seinem Weg vom Anfänger zum Experten als auch in unterschiedlichen Kontexten einen konsistenten und erwartungskonformen Interaktionsmodus bieten soll. Zum anderen ermöglicht ODIN die Exploration und aktive Nutzung von Metadaten über die Heterogenität in einem Informationssystem und integriert unterschiedliche Verfahren zur Heterogenitätsbehandlung mit den informationellen Prozessen des Nutzers.

Cooperative research networks

Jürgen Güdler
Cooperative research networks
Development, structure, and impact in the field of sociology

Bonn: Informationszentrum Sozialwissenschaften 2003
(Research reports; Volume 5),
ISSN 1431-5114; ISBN 3-8206-0140-6

Cooperation in networks is regarded as a special characteristic of modern science. However, it is rare to find studies which – from the viewpoint of development – might substantiate a connection between the professionalization of a discipline and the increase in cooperative action or – from structural aspects – between cooperation in networks and the quality of research. In the study presented here, German sociology is used as an example to examine corresponding relations based on an extensive data set. The databases of the Social Science Information Center form the foundation in this respect. The study population is represented by just under one thousand sociologists who were identified in SOFIS (social science research projects) and pursued a career as a researcher in the late 1970s and early 1980s. Networking is determined on the basis of joint participation in projects and on co-authorships (SOLIS literature database). Statements on the individual advantages of early cooperative action in the form of a later position in the science system are also made through reference to the LEHRE database (social science lectures (discontinued in 1995)). With regard to the question of the long-term development of cooperative action, reference is made at the beginning to other sources which facilitate a comparison between German cooperation trends (1945 to 1995) and American cooperation trends (1895 to 1995).

This study breaks new methodical ground by combining traditional bibliometric methods and actor-oriented linking of data from a wide range of sources for the purpose of structural and biographical analysis. Newly developed methods for measuring cooperation and networking, productivity, research quality, mobility and later placement make a major contribution towards indicator research.

GERSHWIN – A concept of an adaptive help system for software systems

Bernhard Stüber
GERSHWIN – A concept of an adaptive help system for software systems

Bonn: Informationszentrum Sozialwissenschaften 2003
(Reseearch reports; volume 4),
ISSN 1431-5114; ISBN 3-8206-0138-4

In spite of their advantages compared with conventional hypertext help systems, adaptive help systems are very rarely used in commercial software products. This is mainly due to the lack of a generally applicable concept for designing adaptive help systems and missing software tools which support the development of help systems based on such a concept.

GERSHWIN, a concept for the development of an adaptive help system, is presented as a solution to this problem. The research report describes a procedure for developing an adaptive help system and also provides a program library for implementing this procedure. The precondition for the implementation of the concept is access to the source code of the application which should embody an adaptive GERSHWIN help system. If a help system for the application already exists, it is replaced by the GERSHWIN help system. In addition to a visual plan editor for the development of a knowledgebase of plans, the GERSHWIN program library contains all the functions needed for plan recognition and visualization of the help information. The interface between the GERSHWIN help system and the application is based on just four functions in the GERSHWIN program library, thus making it easy to integrate in an application.

The GERSHWIN concept is the result of a cooperation between the GESIS-department Social Science Information Centre (GESIS-IZ), Bonn, and the University of Koblenz-Landau. The research report was accepted as a dissertation at the University of Koblenz-Landau.

Visualization in Document Retrieval

Maximilian Eibl
Visualization in Document Retrieval
Theoretical and practical combination of software ergonomics and graphic design

Bonn: Informationszentrum Sozialwissenschaften 2000
(Research report; Volume 3)
ISSN 1431-5114; ISBN 3-8206-0131-7

Software ergonomics (HC I) and graphic design are two separate schools that have few common goals in the area of designing user interfaces. Whereas ergonomics places the accent on the most effective operation, interface and media design emphasize the aesthetic aspect. The design of a user interface presents the choice between the effective ugliness of ergonomics, and the explorative grace of graphic and media design.

An example for the successful integration of both schools is DEViD (graphic design and software ergonomics integrating visualization for document retrieval systems). DEViD supports users in searching databases and integrates Boolean retrieval as well as probabilistic ranking and vague retrieval within one single interface. DEViD is designed according to aspects of HCI as well as aesthetics. This research report describes how the two schools HCl and graphic design can be combined theoretically as well as practically to develop user interfaces.

DEViD is the result of a cooperation between the Social Science Information Centre, Bonn, the University of Koblenz, and the College of Design, Offenbach. This work was submitted as a PhD-thesis at the University of Koblenz.

The enclosed CD-ROM includes colour versions of the figures, an English and a German animation which demonstrate the usage of DEViD, and a30-page English summary of the report.

Market Analysis and Market Prognosis. The ZVEI Information System ELVIRA

Hansjörg Haas, Jürgen Krause, Jürg Lindlbauer, Thomas Mandl, Bernd Pauer, Ulrich Scheinost, Maximilian Stempfhuber
Market Analysis and Market Prognosis. The ZVEI Information System ELVIRA

Bonn: Informationszentrum Sozialwissenschaften 1998
(Research report; Volume 2)
ISSN 1431-5114
ISBN 3-8206-0126-0

The usage of external information for mangement in small to medium-sized enterprises is far from satisfying. In order to improve this, it is not sufficient to put information online and provide complex query interfaces that require cost-intensive end-user training.
The acceptance of an information system depends on data qualitiy and processing combined with an user-friendly and task-oriented interface suitable for both the casual and the advanced user. Detailed analysis of the domain and information needs of users together with the application of the most recent results in human-computer interaction led to ELVIRA, an electronic information retrieval and analysis system tailored to the needs of industrial associations like the German Electrical and Electronics Manufacturers’ Association (ZVEI), Frankfurt/Main.

ELVIRA is the result of a co-operation between the ZVEI, the German Institute for Economic Research (DIW), Berlin, the ifo-Institute for Economic Research, Munich, and the Social Science Information Centre, Bonn, and is supposed to serve as a model for other trade associations and institutions with similar obligations or ambitions with regard to information.. The project ELVIRA was funded by the Federal Ministry of Economics (grant number II C7-003060/10).

Dynamics of Media Research

Jürgen Güdler
Dynamics of Media Research
A scientometric analysis based on specialized social science databases

Bonn: Informationszentrum Sozialwissenschaften 1996
(Research report; Volume 1)
ISSN 1431-5114
ISBN 3-8206-0117-1

The study investigates thematic trends in the development of German-speaking media research based on more than 9.000 references to media research projects and publications from 1987 through 1994. Applying newly developed, network-analytical measuring procedures, the study develops - for the purposes of a prospective evaluation - thoughts on the future development potential of certain research fields. New types of visualization techniques facilitate the insight into complex structural developments. This research report presented in a highly readable style is addressed to professionals in the fields of media research and science studies.