1. Title of Database: LED display domain + 17 irrelevant attributes

2. Sources:
   (a) Breiman,L., Friedman,J.H., Olshen,R.A., & Stone,C.J. (1984). 
       Classification and Regression Trees.  Wadsworth International
       Group: Belmont, California.  (see pages 43-49).
   (b) Donor: David Aha 
   (c) Date: 11/10/1988

3. Past Usage: (many)
     1. CART book (above):
        -- Optimal Bayes classification rate: 74%
        -- CART decision tree algorithm: 70%
           -- 200 training and 5000 test instances
        -- nearest neighbor algorithm: 41%
     2. Aha,D.W., & Kibler,D. (1988). Unpublished data.
        -- NTgrowth+ instance-based learning algorithm: (500 test instances)
           -- 700 training instances: 70.7%
           -- 1000 training instances: 71.5%

4. Relevant Information Paragraph:
     See the file names create-led.names.  This is an extension of the
     LED display problem, but an additional 17 irrelevant attributes
     are added to the instance space.  Their values are randomly
     assigned the values 0 or 1.  Not an easy problem!

5. Number of Instances: chosen by the user.

6. Number of Attributes: 24 (all Boolean-valued)

7. Attribute Information:
   -- All attribute values are either 0 or 1, according to whether
      the corresponding light is on or not for the decimal digit.
   -- Each attribute (excluding the class attribute, which is an
      integer ranging between 0 and 9 inclusive) has a 10% percent
      chance of being inverted.

8. Missing Attribute Values: None

9. Class Distribution: 10% (Theoretical)
   -- Each concept (digit) has the same theoretical probability
      distribution.  The program randomly selects the attribute.