UCI Machine Learning Repository

["Patrick M. Murphy" <pmurphy@ruffles.ICS.UCI.EDU>]

Message-ID:  <9512061526.aa15955@paris.ics.uci.edu>


Below is a list of databases that have recently been added to the 
UCI Machine Learning Repository.

Any comments or donations would be greatly appreciated
(ml-repository@ics.uci.edu).

Patrick M. Murphy (Librarian)


P.S.

  As I hope to graduate soon and will be starting a full-time
  job come the beginning of the year, this is likely to be my last
  posting regarding the repository.  Mike Pazzani has agreed to
  temporarily take over the day-to-day operations of repository
  maintenance until a new person comes on aboard to take the job
  permanently.  

  It has been an interesting five years communicating with all 
  those people from all over the world who use and donate to the
  repository.  Remember, we are always interested in getting new
  donations!

  Good luck,

  - Patrick




- Peter Turney's cost data used in a recent JAIR paper

  http://ai.iit.nrc.ca/cgi-bin/jair-abstract?turney95a

  Cost data is available for each of the following databases:
  ann-thyroid, bupa-liver, heart-disease, hepatitis, and 
  pima-indians-diabetes.  All sets of cost data are in separate
  costs/ directories that are associated with their respective 
  databases.

- vowel-context data (donated by Peter Turney)

  An extension of connectionist-bench vowel data.

  In my work on context-sensitive learning, I used the "Deterding 
  Vowel Recognition Data", but I found it necessary to reformulate 
  the data.  Implicit in the original data is contextual information 
  on the speaker's gender and identity. For my work, it was necessary 
  to make this information explicit. The file "vowel-context.data" 
  adds the speaker's sex and identity as new features. The format of 
  the data file is described below. -- Turney

  Located in undocumented/connectionist-bench/vowel/

- Mobile Robots (donated by Klingspor, Morik and Rieger)

  Learning concepts from sensor data of a mobile robot.

  We provide here a set of data sets, where each data set corresponds 
  to learning disjoint concepts at one level. The levels are organized 
  in a hierarchy ...  a sequence of learning passes can learn high-level 
  concepts from raw sensor data... 

- East-West Challenge data and results (donated by Peter Turney)

  These files describe the competition and contain all of the material
  that was made available to the competitors, the algorithms used and 
  the results of the competition.

  Located in trains/