Re: DM: Transforming Data

• Organization: Kansas State University

Eni,

     Here are some suggestions:

1. Chapters 4, 6, and 7 of

    Predictive Data Mining: A Practical Guide
    S. M. Weiss and N. Indurkhya
    Book page: http://www.mkp.com/books_catalog/1-55860-533-9.asp
    Software page: http://www.data-miner.com

    also have some material on data reduction and its effects
    on later stages of the pipeline.  I just started reviewing
    this book for a summer DM course I'm preparing, and it
    seems to address important issues in this vein.

2. I think the KDD overview paper by Fayyad (in the Fayyad,
    Piatetsky-Shapiro, and Uthurusamy book from AAAI/MIT
    Press) addresses "downstream effects" of transformation
    and reduction as well.

    "Knowledge Discovery and Data Mining: Towards a
    Unifying Framework", Usama Fayyad, Gregory Piatetsky-Shapiro,
    and Padhraic Smyth. Proceedings of Second International
    Conference on Knowledge Discovery and Data Mining
    (KDD-96), AAAI Press, 1996.
    http://www.research.microsoft.com/~fayyad/papers/kdd96-intro.htm

3. We just published an MLJ paper addressing problem reformulation
    by attribute partitioning, and its effects on unsupervised learning
    (cluster definition) and model selection in subsequent supervised
    learning.  The paper is geared towards time series learning, but we
    have used the technique for other KDD applications.

    A Multistrategy Approach to Classifier Learning from Time Series.
    W. H. Hsu, S. R. Ray, and D. C. Wilkins.  Machine Learning,
    38:213-236, 2000.
    Preprint:
http://lightsaber.ncsa.uiuc.edu/Papers/Hsu/mlmsl-final-kluwer.doc
    Related papers/thesis: http://www.cis.ksu.edu/~bhsu/

Hope this helps,
Bill

----- Original Message -----
From: "Dale Jelinek" <dalejelinek@hotmail.com>
To: <datamine-l@nautilus-sys.com>
Sent: Sunday, April 16, 2000 2:39 PM
Subject: Re: DM: Transforming Data


 >
 > Check out a book called "Data Preparation for Data Mining" by Dorian Pyle.
 > It is available from Amazon.com
 >
 > ----Original Message Follows----
 > From: Eniana Mustafaraj <eni@informatik.uni-siegen.de>
 > Reply-To: datamine-l@nautilus-sys.com
 > To: datamine-l@nautilus-sys.com
 > Subject: DM: Transforming Data
 > Date: Fri, 14 Apr 2000 13:52:09 +0200 (CEST)
 >
 > Dear friends,
 >
 > could you help me with some literature suggestions in the topics:
 >
 > - effects of transforming the data
 > - impact of the loss of information in the further analysis of the data
 >
 >
 > Thank you,
 >
 > Eni
 >
 >
 > ______________________________________________________
 > Get Your Private, Free Email at http://www.hotmail.com

=======================================================
  William H. Hsu, Ph.D.
  Director, KDD Lab, Kansas State University
  Research Scientist, Automated Learning Group, NCSA
  bhsu@cis.ksu.edu, bhsu@ncsa.uiuc.edu
  http://www.cis.ksu.edu/~bhsu           ICQ: 28651394
=======================================================

• References:
  • Re: DM: Transforming Data
    • From: Dale Jelinek

• Prev by Date: Re: DM: Transforming Data
• Next by Date: DM: Modern Regression and Classification Course- Washington
• Prev by thread: Re: DM: Transforming Data
• Next by thread: DM: Datamining Definition
• Index(es):
  • Date
  • Thread