About DIAsDEM

DIAsDEM is a research project that deals with the integration of legacy data and semi-structured documents by employing data mining techniques. The project is financed by the German Research Society (Deutsche Forschungsgesellschaft, DFG). DIAsDEM consists of two research groups: the team of Prof. Dr. Stefan Conrad (Heinrich Heine University, Düsseldorf) and the team of Prof. Dr. Myra Spiliopoulou (Otto-von-Guericke-University, Magdeburg).

DIAsDEM aims at using information being hidden in unstructured data in order to support sense-making processes. This information is mostly included in insufficiently documented data sources. Some of these sources are erroneous, incomplete or contradictory. Data mining is used to comprehend the semantics of this information and to consolidate them in an information system of integrated and reliable data for further analyses and enquiries. In this connection tasks are in particular as follows:

# Detection of semantic characteristics of data and possibly existent dependencies between semantically related data.

# Utilisation of discovered semantics for the integration of data in a homogeneous information system.

Prof. Dr. Spiliopoulou's team is mainly dealing with the extraction of semantic characteristics from application-specific text archives by methods of data mining. At this stage, more information concerning DIAsDEM can either be found in the project application (Postscript-File, PDF-File) or in our recent publications.

Team Magdeburg:

• Prof. Dr. Myra Spiliopoulou (Head of Project)
• Henner Graubitz (Research Assistant)
• Karsten Winkler (Research Assistant)

Team Düsseldorf

• Prof. Dr. Stefan Conrad (Head of Project)
• Evguenia Altareva (Research Assistant)

Download:

• Check out the newest version of the DIAsDEM-Workbench via Sourceforge
• Check out the newest information from HypKNOWsys

Online Demo

• We prepared also a flash animated Demonstration. You can use it also to get an impression about the DIAsDEM-Workbench and the research behind.

Publications

Conference Proceedings:

• Henner Graubitz and Myra Spiliopoulou. Deriving Multiple Topics to Label Small Document Regions. Proceedings of the 6th International Conference on Data Warehousing and Knowledge Discovery (DaWaK) Zaragoza, Spain, September 2004
• Myra Spiliopoulou, Anja Schulz und Karsten Winkler. Text Mining an der Börse: Einfluss von Ad-hoc-Mitteilungen auf die Kursentwicklung. In Christine Becker und Heinz Redlich (Hrsg.): Data Mining und Statistik in Hochschule und Wirtschaft, Proceedings der 7. Konferenz der SAS-Anwender in Forschung und Entwicklung (KSFE), 20. Februar - 21. Februar 2003, Universität Potsdam, S. 215-228. Shaker Verlag, Aachen, 2003. PDF-File

• Karsten Winkler and Myra Spiliopoulou. Text Mining in der Wettbewerberanalyse: Konvertierung von Textarchiven in XML-Dokumente. In Proceedings der 6. Konferenz der SAS Anwender in Forschung und Entwicklung, pages 347-363, Shaker Verlag, Aachen, Germany, 2003. Download: Postscript-Fil e, PDF-File

• Karsten Winkler and Myra Spiliopoulou. Structuring Domain-Specific Text Archives by Deriving a Probabilistic XML DTD. In Proceedings of the 6th European Conference on Principles and Practice of Knowledge Discovery in Databases (PKDD'02), pages 461-474, Helsinki, Finland, August 2002. Springer-Verlag, Berlin, Heidelberg.

• Karsten Winkler and Myra Spiliopoulou. Semi-Automated XML Tagging of Public Text Archives: A Case Study. In Proceedings of EuroWeb 2001 "The Web in Public Administration", pages 271-285, Pisa, Italy, December 2001. Download: Postscript-File, PDF-File

• Karsten Winkler and Myra Spiliopoulou. Extraction of Semantic XML DTDs from Texts Using Data Mining Techniques. In Proceedings of the K-CAP 2001 Workshop on Knowledge Markup and Semantic Annotation, pages 59-68, Victoria, BC, Canada, October 2001. Download: Postscript-File, PDF-File

• Karsten Winkler and Myra Spiliopoulou. Integrating Data and Probabilistically Structured Text Documents. In Andreas Bauer, Susanne Busse, Ralf-Detlef Kutsche and Wolfgang Lehner, editors, Proceedings des 5. Workshops "Förderierte Datenbanken" und GI Arbeitstreffen "Konzepte des Data Warehousing" (FDBS 2001), pages 16-29, Berlin, Germany, October 2001. Forschungsberichte des Fachbereichs Informatik, Bericht-Nr. 2001-14, Technische Universität Berlin. Download: Postscript-File, PDF-File

• Henner Graubitz, Myra Spiliopoulou and Karsten Winkler. The DIAsDEM Framework for Converting Domain-Specific Texts into XML Documents with Data Mining Techniques. In Proceedings of the First IEEE International Conference on Data Mining, pages 171-178, San Jose, CA, USA, November / December 2001. IEEE Computer Society, Los Alamitos.
• Henner Graubitz, Karsten Winkler and Myra Spiliopoulou. Semantic Tagging of Domain-Specific Text Documents with DIAsDEM. In Kai-Uwe Sattler, editor, Proceedings of the 1st International Workshop on Databases, Documents, and Information Fusion (DBFusion 2001), pages 61-72, Gommern, Germany, May 2001. Preprint Nr. 8, Institut für Informatik, Otto-von-Guericke-Universität Magdeburg. Download: Postscript-File, PDF-File

Chapters in Books:

• Anja Schulz, Myra Spiliopoulou und Karsten Winkler. Kursrelevanzprognose von Ad-hoc-Meldungen: Text Mining wider die Informationsüberlastung im Mobile Banking. In: Wolfgang Uhr, Werner Esswein und Eric Schoop (Hrsg.). Wirtschaftsinformatik 2003: Medien, Märkte, Mobilität, Band II, S. 181-200. Physika-Verlag, Heidelberg, 2003. Download: PDF-File

• Myra Spiliopoulou and Karsten Winkler. Text Mining auf Handelsregistereinträgen: Der SAS Enterprise Miner im Einsatz. In Klaus D. Wilde, Hajo Hippner, Melanie Merzenich, editors, Data Mining: Mehr Gewinn aus Ihren Kundendaten, pages 117-124. Verlagsgruppe Handelsblatt, Düsseldorf, 2002. Download: Postscript-File , PDF-File

Doctoral Consortia:

• Karsten Winkler. Semantic Tagging of Domain-Specific Text Archives with Data Mining Techniques. In Doctoral Consortium at ECIS 2002: The Xth European Conference on Information Systems, Gdansk, Poland, June 2002. URL: http://ecis2002.univ.gda.pl/ecis2002.html?s=phd. Download: Postscript-File, PDF-File

• Karsten Winkler. Semantische Auszeichnung von Textarchiven mit Data Mining-Methoden. In Doctorial Consortium im Vorfeld der WI-IF 2001, Schloß Reisensburg bei Günzburg, Germany, September 2001. URL: http://wi.oec.uni-bayreuth.de/doctoral/. Download: Postscript-File, PDF-File

Workgroup:

UMD

Institute

Research Group

Research

Staff

News

Library