Exploratory data analysis using data semantics

  • Friedrich Gebhardt


In our project EXPLORA, we try to utilize the semantics of the data for their exploratory statistical interpretation. The kernel of the system supplies a set of tools built around a search algorithm that exploits the semantic structures among variables, variable values and other objects. When a data set (or a type of data sets) is set up, these objects and relations have to be initialized; this includes defining the evaluation methods applicable to this particular data set. The system finds the most interesting statements about the data and thereby suppresses statements that are redundant or uninteresting relative to other ones that have been displayed. We propose an algorithm working on the data objects covered by a statement rather than on its syntactical form for suppressing results that apply to nearly the same data objects as another one even though the syntactical forms may — or may not — be quite different.


Search Space Search Algorithm Expert System Data Object Exploratory Data Analysis 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. [1]
    Chowdhury, Shamsul I.: Statistical expert systems — a special application area for knowledge—based computer methodology. Linköping: University, 1987.109 pp.Google Scholar
  2. [2]
    Ellman, Thomas: Explanation-based learning: a survey of programs and perspectives. In: Computing Surveys 21 (1989), pp. 163 – 221.Google Scholar
  3. [3]
    Gale, William A.: Statistical applications of artificial intelligence and knowledge engineering. In: Knowledge Engineering Review 2 (1988), pp. 227 – 247.Google Scholar
  4. [4]
    Gebhardt, Friedrich: Prospects for expert systems to analyze election data. In: Classification and related methods of data analysis: proceedings of the First Conference of the International Federation of Classification Societies (IFCS) (Aachen, Juni 1987)/Hans H. Bock (ed.). Amsterdam: North-Holland, 1988, pp. 691 – 696.Google Scholar
  5. [5]
    Gebhardt, Friedrich: Statistische Fragestellungen bei einem Expertensystem zur explorativen Datenanalyse. Sankt Augustin: GMD, 1988 (GMD-Studien 137). – 123 pp.Google Scholar
  6. [6]
    Gebhardt, Friedrich: An expert system strategy for selecting interesting results. In: COMPSTAT 90: proceedings (Dubrovnik, Sept. 1990). Heidelberg: Physica, 1990. To be printed.Google Scholar
  7. [7]
    Gebhardt, Friedrich: Choosing among competing generalizations. Sankt Augustin: GMD, 1989 (Arbeitspapiere der GMD, 421). — 16 pp. — Preprint, to be published elsewhere.Google Scholar
  8. [8]
    Gebhardt, Friedrich: Choice among overlapping classifications. Submitted to the proceedings of 14. Jahrestagung der Gesellschaft für Klassifikation (Marburg, March 1990)/P. Ihm (ed.); H.-H. Bock (ed.). Berlin: Springer, 1990.Google Scholar
  9. [9]
    Hand, David J.: Expert systems in statistics. In: Knowledge Engineering Review 1, Nr. 3 (1986), pp. 2 — 10.Google Scholar
  10. [10]
    and, David J.: The application of expert systems in statistics. In: Interactions in artificial intelligence and statistical methods/Bob Phelps (cd.). Aldershot: Gower Technical Press, 1987, pp. 3 — 17.Google Scholar
  11. [11]
    Hoschka, Peter; Klösgen, Willi: A support system for interpreting statistical data. In: Knowledge discovery in databases/G. Piatetsky-Shapiro (ed.); W. Frawley (ed.). Cambridge, Mass.: MIT-Press. — In preparation, 1990.Google Scholar
  12. [12]
    Klösgen, Willi: The generalization step in a statistics interpreter. In: Data Analysis, Leaming Symbolic and Numeric Knowledge/E. Diday (cd.). New York: Nova Science, 1989, pp. 473–480.Google Scholar
  13. [13]
    Klösgen, Willi: EXPLORA: content interpretation of statistical data. In: Fortschritte der Statistik-Software/F. Faulbaum (ed.); H.M. Kehlinger (ed.). Stuttgart: G. Fischer, 1990. In print.Google Scholar
  14. [14]
    Latocha, Peter: Searching for the most interesting facts. In: Proceedings of the TASTED International Symposium on Expert Systems: theory and applications. Zürich: Acta Press, 1989, pp. 90 — 93.Google Scholar
  15. [15]
    Latocha, Peter: Exploration von Aussagenräumen: ein semantischer Ansatz. Sankt Augustin: GMD, 1989 (GMD-Studien 164). — 144 pp.Google Scholar
  16. [16]
    Wittkowski, Knut M.: Knowledge based support for the management of statistical databases. In: Statistical and scientific database management: 4th International Working Conference SSDBM (Rom, Juni 1988)/M. Rafanelli (cd.); J.C. Klensin (ed.); P. Svensson (ed.). Berlin: Springer, 1989, pp. 62 — 71.Google Scholar

Copyright information

© Springer-Verlag/Wien 1990

Authors and Affiliations

  • Friedrich Gebhardt
    • 1
  1. 1.Sankt Augustin 1Deutschland

Personalised recommendations