Constructing Efficient Decision Trees by Using Optimized Numeric Association Rules.

@inproceedings{DBLP:conf/vldb/FukudaMMT96,
  author    = {Takeshi Fukuda and
               Yasuhiko Morimoto and
               Shinichi Morishita and
               Takeshi Tokuyama},
  editor    = {T. M. Vijayaraman and
               Alejandro P. Buchmann and
               C. Mohan and
               Nandlal L. Sarda},
  title     = {Constructing Efficient Decision Trees by Using Optimized Numeric
               Association Rules},
  booktitle = {VLDB'96, Proceedings of 22th International Conference on Very
               Large Data Bases, September 3-6, 1996, Mumbai (Bombay), India},
  publisher = {Morgan Kaufmann},
  year      = {1996},
  isbn      = {1-55860-382-4},
  pages     = {146-155},
  ee        = {db/conf/vldb/FukudaMMT96.html},
  crossref  = {DBLP:conf/vldb/96},
  bibsource = {DBLP, http://dblp.uni-trier.de}
}

Abstract

We propose an extension of an entropy-based heuristic of Quinlan [Q93] for constructing a decision tree from a large database with many numeric attributes. Quinlan pointed out that his original method (as well as other existing methods) may be inefficient if any numeric attributes are strongly correlated. Our approach offers one solution to this problem. For each pair of numeric attributes with strong correlation, we compute a two-dimensional association rule with respect to these attributes and the objective attribute of the decision tree. In particular, we consider a family R of grid-regions in the plane associated with the pair of attributes. For R in R, the data can be split into two classes: data inside R and data outside R.

We compute the region Ropt in R that minimizes the entropy of the splitting, and add the splitting associated with Ropt (for each pair of strongly correlated attributes) to the set of candidate tests in Quinlan's entropy-based heuristic.

We give efficient algorithms for cases in which R is (1) x-monotone connected regions, (2) based-monotone regions, (3) rectangles, and (4) rectilinear convex regions. The algorithm for the first case has been implemented as a subsystem of SONAR(System for Optimized Numeric Association Rules) developed by the authors. Tests show that our approach can create small-sized decision trees.

Copyright © 1996 by the VLDB Endowment. Permission to copy without fee all or part of this material is granted provided that the copies are not made or distributed for direct commercial advantage, the VLDB copyright notice and the title of the publication and its date appear, and notice is given that copying is by the permission of the Very Large Data Base Endowment. To copy otherwise, or to republish, requires a fee and/or special permission from the Endowment.

Online Paper

• Download PDF file (www.vldb.org, Darmstadt, Germany)
• Download PDF file (www.acm.org, New York, USA)

ACM SIGMOD Anthology

• Windows: Click the letter of your CD drive
  A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
• Mac: Click here
• UNIX/LINUX: mount the CD and click on the path of your mount point:
  /Anthology/vldb8997 or /cdrom

• Windows: Click the letter of your CD drive
  A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
• Mac: Click here
• UNIX/LINUX: mount the DVD and click on the path of your mount point:
  /Anthology/aDVD1 or /dvd

Printed Edition

Electronic Edition

• From SunSITE Central Europe (Aachen, Germany)
  • Postscript file, via HTTP
  • Sources, via HTTP
• From CS Dept., University Trier (Germany)
  • Postscript file, via HTTP
  • Postscript file, via anonymous FTP
  • Postscript file, compressed by gzip, via anonymous FTP
  • Postscript file, compressed by compress, via anonymous FTP
  • Sources, via HTTP
  • Sources, via anonymous FTP

References

[ACKT96]

Tetsuo Asano, Danny Z. Chen, Naoki Katoh, Takeshi Tokuyama: Polynomial-Time Solutions to Image Segmentation. SODA 1996: 104-113

[AGL+92]

Rakesh Agrawal, Sakti P. Ghosh, Tomasz Imielinski, Balakrishna R. Iyer, Arun N. Swami: An Interval Classifier for Database Mining Applications. VLDB 1992: 560-573

[ALS93]

Rakesh Agrawal, Tomasz Imielinski, Arun N. Swami: Database Mining: A Performance Perspective. IEEE Trans. Knowl. Data Eng. 5(6): 914-925(1993)

[AT94]

...

[BFOS84]

...

[CG86]

Bernard Chazelle, Leonidas J. Guibas: Fractional Cascading: I. A Data Structuring Technique. Algorithmica 1(2): 133-162(1986)

[DE93]

David P. Dobkin, David Eppstein: Computing the Discrepancy. Symposium on Computational Geometry 1993: 47-52

[DEY86]

David P. Dobkin, Herbert Edelsbrunner, Chee-Keng Yap: Probing Convex Polytopes. STOC 1986: 424-432

[FMMT96a]

Takeshi Fukuda, Yasuhiko Morimoto, Shinichi Morishita, Takeshi Tokuyama: Mining Optimized Association Rules for Numeric Attributes. PODS 1996: 182-191

[FMMT96b]

Takeshi Fukuda, Yasuhiko Morimoto, Shinichi Morishita, Takeshi Tokuyama: Data Mining Using Two-Dimensional Optimized Accociation Rules: Scheme, Algorithms, and Visualization. SIGMOD Conference 1996: 13-23

[FMMY96c]

Takeshi Fukuda, Yasuhiko Morimoto, Shinichi Morishita, Takeshi Tokuyama: SONAR: System for Optimized Numeric AssociationRules. SIGMOD Conference 1996: 553

[GJ79]

M. R. Garey, David S. Johnson: Computers and Intractability: A Guide to the Theory of NP-Completeness. W. H. Freeman 1979, ISBN 0-7167-1044-7

[HR76]

Laurent Hyafil, Ronald L. Rivest: Constructing Optimal Binary Decision Trees is NP-Complete. Inf. Process. Lett. 5(1): 15-17(1976)

[MAR96]

Manish Mehta, Rakesh Agrawal, Jorma Rissanen: SLIQ: A Fast Scalable Classifier for Data Mining. EDBT 1996: 18-32

[MST84]

...

[PR90]

...

[Q86]

J. Ross Quinlan: Induction of Decision Trees. Machine Learning 1(1): 81-106(1986)

[Q93]

J. Ross Quinlan: C4.5: Programs for Machine Learning. Morgan Kaufmann 1993, ISBN 1-55860-238-0

[QR89]

J. Ross Quinlan, Ronald L. Rivest: Inferring Decision Trees Using the Minimum Description Length Principle. Inf. Comput. 80(3): 227-248(1989)