Re: DM: Standard test suites/benchmarking

[ragrawal]

Checkout the Quest website:

              http://www.almaden.ibm.com/cs/quest

for some synthetic data generation programs.

Cheers

/rakesh





ejs2c@watt.seas.virginia.edu on 11/03/97 10:36:49 AM

Please respond to ejs2c@watt.seas.virginia.edu

To:   datamine-l@nautilus-sys.com
cc:    (bcc: Rakesh Agrawal/Almaden/IBM)
Subject:  DM: Standard test suites/benchmarking






Hi folks... I'm looking for some advice,
I am in currently working on an undergraduate thesis project dealing 
with
the evaluation of data mining tools.  To evaluate these tools I would 
be
interested to know if there is any collection of standard test suites
(data sets) which might be used.  I have found the collection of data 
sets
at www.kdnuggets.com.  This listing is great, but overwhelming - It is
very hard to tell which of these data sets would be useful for 
evaluating
software packages.
I would appreciate any advice on which data sets to use and where
to find these sets.  The problems could be simple or complex, but I am
specifically interested in data sets which are representative of
certain "types" of statistical problems, or data sets which are
considered classic examples.  Also, if I am looking in the wrong 
direction
here, let me know as well.
Please respond with any hints or advice you might have on the data 
sets I
might use for testing data mining tools.
Thank You in Advance,
Eric Schmidt
University of Virginia
Systems Engineering
Class of 1998