Abstract
This paper analyses properties of conceptual hierarchy obtained via incremental concept formation method called “flexible prediction” in order to determine what kind of “relevance” of participating attributes may be requested for meaningful conceptual hierarchy. The impact of selection of simple and combined attributes, of scaling and of distribution of individual attributes and of correlation strengths among them is investigated.
Paradoxically, both: attributes weakly and strongly related with other attributes have deteriorating impact onto the overall classification. Proper construction of derived attributes as well as selection of scaling of individual attributes strongly influences the obtained concept hierarchy. Attribute density of distribution seems to influence the classification weakly
It seems also, that concept hierarchies (taxonomies) reflect a compromise between the data and our interests in some objective truth about the data.
To obtain classifications more suitable for one’s purposes, concept hierarchy of attributes rather than of objects should be looked for. A proposal of bayesian network based clustering is proposed.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
C.K.Chow, C.N.Liu: Approximating discrete probability distributions with dependence trees, IEEE Transactions on Information Theory, Vol. IT-14, No. 3, (1968), 462–467
Fisher D., Langley P.: Approaches to conceptual clustering, Proc. 9th IJCAI, Los Angeles, 1985, pp. 691–697
Fisher D.: Knowledge acquisition via incremental conceptual clustering, Machine Learning 2,2, 1987, pp. 139–172
Fisher D.H.: Conceptual clustering. learning from examples and inference, Proc. 4th International Workshop on Machine Learning, Irvine, Morgan Kaufman, 1987, pp. 38–49
Fisher D.H.: Noise-tolerant conceptual clustering, Proc. IJCAI’89 Vol. 2, pp. 825–830
Fisher, D., Hapanyengwi, G. “Database Management and Analysis Tools of Machine Learning.” Journal of Intelligent Information Systems, 2, 1993, 5–38.
Fisher, D.. Iterative optimization and simplification of hierarchical clusterings. Journal of Artificial Intelligence Research, 4, 1996, 147–178.
Gennari J.H., Langley P., Fisher D.: Models of incremental concept formation, Artificial Intelligence 40 (1989) 11–61
Klopotek M.A.: On the Phenomenon of Flattening ‘Flexible Prediction’ Concept Hierarchy, in: Ph. Jorrand, J. Kelemen, Eds.: Fundamentals of Artificial Intelligence Research, International Workshop Smolenice Czech-Slovakia, 8–13 Sept. 1991, Lecture Notes in Artificial Intelligence 535, Springer-Verlag, Berlin Heidelberg New York 1991, 99–111
Klopotek M.A.: Dependence of evaluation function on correlation coefficient in the concept formation method “FLEXIBLE PREDICTION” (in Polish) in: P. Sienkiewicz, J. Tchórzewski Eds.: Sztuczna Inteligencja i Cybernetyka Wiedzy (cybernetyka–inteligencja–rozwój), PTC, WSRP w Siedlcach. SiedlceWarszawa 23–24.9. 1991, pp. 37–42
M.A.Klopotek, A.Matuszewski: On Irrelevance of Attributes in Flexible Prediction. Proc. 2nd Int. Conf. on New Techniques and Technologies for Statistics (NTTS’95), Bonn, 19–22 Nov., 1995, Publisher: GMD Sankt Augustin, pp 282–293.
P.Spirtes, C.Glymour, R.Scheines: Causation, Prediction and Search, Lecture Notes in Statistics 81, Springer-Verlag, 1993.
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2001 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Klopotek, M.A. (2001). Taxonomy Building: Cases or Attributes ?. In: Kłopotek, M.A., Michalewicz, M., Wierzchoń, S.T. (eds) Intelligent Information Systems 2001. Advances in Intelligent and Soft Computing, vol 10. Physica, Heidelberg. https://doi.org/10.1007/978-3-7908-1813-0_9
Download citation
DOI: https://doi.org/10.1007/978-3-7908-1813-0_9
Publisher Name: Physica, Heidelberg
Print ISBN: 978-3-7908-1407-1
Online ISBN: 978-3-7908-1813-0
eBook Packages: Springer Book Archive