Learning of Class Descriptions from Class Discriminations: A Hybrid Approach for Relational Objects

  • Peter Geibel
  • Kristina Schädler
  • Fritz Wysotzki
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 2479)


The paper addresses the question how learning class discrimination and learning characteristic class descriptions can be related in relational learning. We present the approach TRITOP/MATCHBOX combining the relational decision tree algorithm TRITOP with the connectionist approach MATCHBOX. TRITOP constructs efficiently a relational decision tree for the fast discrimination of classes of relational descriptions, while MATCHBOX is used for constructing class prototypes.

Although TRITOP’s decision trees perform very well in the classification task, they are difficult to understand and to explain. In order to overcome this disadvantage of decision trees in general, in a second step the decision tree is supplemented by prototypes. Prototypes are generalised graphtheoretic descriptions of common substructures of those subclasses of the training set that are defined by the leaves of the decision tree. Such prototypes give a comprehensive and understandable description of the subclasses. In the prototype construction, the connectionist approach MATCHBOX is used to perform fast graph matching and graph generalisation, which are originally NP-complete tasks.


Decision Tree Node Variable Class Description Inductive Logic Programming Bidirectional Associative Memory 
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.
    G. Bisson. Conceptual clustering in a first order logic representation. In B. Neumann, editor, Proc. of the 10th EC AI, pages 458–462. John Wiley & Sons, 1992.Google Scholar
  2. 2.
    H. Blockeel, L. DeRaedt, and J. Ramon. Top-down induction of clustering trees. In J. Shavlik, editor, Proceedings of the 15th International Conference on Machine Learning, pages 55–63, 1998.Google Scholar
  3. 3.
    J. Bruck and J. W. Goodman. On the power of neural networks for solving hard problems. Journal of Complexity, 6:129–135, 1990.zbMATHCrossRefMathSciNetGoogle Scholar
  4. 4.
    L.I. Burke and J.P. Ignizio. Neural networks and operations research: An overview. Computers Ops. Res., 19(3/4):179–189, 1992.CrossRefzbMATHGoogle Scholar
  5. 5.
    B. Dolsak and S. Muggleton. The application of Inductive Logic Programming to finite element mesh design. In S. Muggleton, editor, Inductive Logic Programming. Academic Press, London, 1992.Google Scholar
  6. 6.
    W. Emde and D. Wettschereck. Relational instance based learning. In L. Saitta, editor, Proc. 13th ICML, pages 122–130. Morgan Kaufmann Publishers, 1996.Google Scholar
  7. 7.
    F. Esposito, A. Laterza, D. Malerba, and G. Semerano. Refinement of datalog programs. In ICML’ 96 Workshop on “Datamining with Inductive Logic Programming, pages 73–94, 1996.Google Scholar
  8. 8.
    J. A. Feldman, M. A. Fanty, N. H. Goddard, and K. J. Lynne. Computing with structured connectionist networks. CACM, 31(2):170–187, February 1988.Google Scholar
  9. 9.
    P. Geibel. Induktive Konstruktion von merkmalsbasierten und logischen Klassifikatoren für relationale Strukturen. PhD thesis, TU Berlin, 1999.Google Scholar
  10. 10.
    P. Geibel and F. Wysotzki. Learning relational concepts with decision trees. In L. Saitta, editor, Machine Learning: Proc. of the 13th Int. Conf. Morgan Kaufmann Publishers, San Fransisco, CA, 1996.Google Scholar
  11. 11.
    P. Geibel and F. Wysotzki. A logical framework for graph theoretic decision tree learning. In N. Lavrac and S. Dzeroski, editors, Proc. ILP 97. Springer, 1997.Google Scholar
  12. 12.
    D. Haussler. Learning Conjunctive Concepts in Structural Domains. Machine Learning, 4:7–40, 1989.Google Scholar
  13. 13.
    F. Hayes-Roth and J. McDermott. Knowledge acquisition from structural descriptions. In Raj Reddy, editor, Proceedings of the 5th International Joint Conference on Artificial Intelligence, pages 356–362, Cambridge, MA, August 1977. William Kaufmann.Google Scholar
  14. 14.
    N. Helft. Inductive generalization: A logical framework. In Proceedings of the Second Working Session on Learning, pages 149–157, 1987.Google Scholar
  15. 15.
    J.-U. Kietz. A comparative study of structural most specific generalisations used in machine learning. In Proc. ILP-93, pages 149–164, Ljubljana, Slovenia, 1993. J. Stefan Institute Tech. Rep. IJS-DP-6707.Google Scholar
  16. 16.
    M. Manago. Knowledge intensive induction. In A. M. Segre, editor, Proceedings of the 6th International Workshop on Machine Learning, pages 151–155, Ithaca, 1989. Morgan Kaufmann.Google Scholar
  17. 17.
    G. D. Plotkin. A note on inductive generalization. In Machine Intelligence, pages 153–164. Edinburgh University Press, 1969.Google Scholar
  18. 18.
    L. De Raedt and H. Blockeel. Using logical decision trees for clustering. In N. Lavrač and S. Džeroski, editors, Proc. of the 7th Int. WS on ILP, volume 1297 of LNAI, pages 133–140. Springer, September 17–20, 1997.Google Scholar
  19. 19.
    K. Schädler and F. Wysotzki. A connectionist approach to distance-based analysis of relational data. In X. Liu, P. Cohen, and M. Berthold, editors, Proc. of the IDA-97, pages 137–148. Springer, 1997.Google Scholar
  20. 20.
    K. Schädler and F. Wysotzki. Theoretical foundations of a special neural net approach for graphmatching. Technical Report 96-26, TU Berlin, CS Dept., 1996.Google Scholar
  21. 21.
    K. Schädler and F. Wysotzki. Comparing structures using a hopfield-style neural network. Applied Intelligence-special issue on Neural Networks and Structured Knowledge, 11(1):5–30, 1999.Google Scholar
  22. 22.
    A. Srinivasan, S. Muggleton, R. King, and M. Sternberg. Mutagenesis: Ilp experiments in a non-determinate biological domain. In S. Wrobel, editor, Proc. of the Fourth Intl. WS on ILP, number 237 in GMD-Studien, pages 217–232, 1994.Google Scholar
  23. 23.
    A. Srinivasas, R.D. Kind, S.H. Muggleton, and M. J. E Sternberg. Carcinogenesis predictions using ILP. In N. Lavrac and S. Dzeroski, editors, Proc. ILP-97, number 1297 in LNAI, pages 273–287. Springer-Verlag, 1997.Google Scholar
  24. 24.
    S. A. Vere. Induction of Concepts in the Predicate Calculus. In Proc. of the Fourth Intl. Joint Conf. on AI, volume 1, pages 281–287, 1975.Google Scholar
  25. 25.
    F. Wysotzki. Artificial Intelligence and Artificial Neural Nets. In Proc. 1st WS on AI, Shanghai, September 1990. TU Berlin and Jiao Tong University Shanghai.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2002

Authors and Affiliations

  • Peter Geibel
    • 1
  • Kristina Schädler
    • 2
  • Fritz Wysotzki
    • 1
  1. 1.Methods of Artificial Intelligence, Computer Science Department, Sekr. Fr 5-8Technical University BerlinBerlinGermany
  2. 2.Pace Aerospace Engineering and Information Technology GmbHBerlin

Personalised recommendations