Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: Notesfiles $Revision: 1.6.2.17 $; site uiucdcsb.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!mhuxr!ihnp4!inuxc!pur-ee!uiucdcsb!jbecker
From: jbecker@uiucdcsb.UUCP
Newsgroups: net.wanted
Subject: Wanted: raw data
Message-ID: <12500016@uiucdcsb.UUCP>
Date: Thu, 24-Jan-85 20:25:00 EST
Article-I.D.: uiucdcsb.12500016
Posted: Thu Jan 24 20:25:00 1985
Date-Received: Sat, 26-Jan-85 06:26:55 EST
Lines: 44
Nf-ID: #N:uiucdcsb:12500016:000:1321
Nf-From: uiucdcsb!jbecker    Jan 24 19:25:00 1985


I need data bases for machine inductions experiments.
The particular domain is unimportant, but the data should have
the following characteristics:

1: There should be several (10 ... 1000) examples for each of several (2 .. 50)
   decision classes.

2: Each example should be characterized in terms of a values for several
   (5 .. 100) attributes.

3: Any format is acceptable, but some machine readable form would be best.

A micro-example:

  class  blood-temp  body-covering  eggs  habitat  milk  weight

  birds      hot        feathers    yes     air     no     1
  birds      hot        feathers    yes     air     no    10

  mammals    hot        fur         no      land    yes  175
  mammals    hot        fur         yes     water   yes   15

  reptiles   cold       scales      yes     land    no     3
  reptiles   cold       scales      yes     water   no    75


From the sets of examples the program will generate a rule describing
each class, which can then be used in classifying new examples (which
may or may not have been in the training set).

If you think you have something I can use, please drop me a line:

	  Jeff Becker
	  154 Digital Computer Lab
	  Department of Computer Science
	  University of Illinois
	  1304 W. Springfield
	  Urbana, IL 61810

	  jbecker@uiucdcs

	  (217) 333-1376