Public CAP Data
We're currently in the process of making the CAP data
available for public distribution.
For a sample, consider the data we put together for a recent
homework assignment in a machine learning
July 1994 Communications of the ACM Paper
Stay tuned to this page for additional developments!
CACM data files
I suggest selecting the mosaic option "load to local disk" before selecting any of the following files.
test.foo Try this to make sure that things work for you.
Compressed to 2meg, uncompresses to 14meg. Don't select this lightly! NOTE: this data is a superset of the CACM paper data.
Compressed to .5meg, uncompresses to 4meg.
User1 CACM data only Compressed to 1meg, uncompresses to 10meg.
We've also converted part of the CACM database into C4.5 format. There
are three files:
Note, there are more columns in this file than listed in the .names
file; the last few columns are class assignments for various experiments.
- cap-kb.names. You'll need to insert the class description at the
top of this file for the particular experiment you want to run.
cap-kb.key, a succinct description of the encoding used in the above
Last updated February 4, 1998, by firstname.lastname@example.org.