Results of the KDD'99 Knowledge Discovery Contest

For the results of the other KDD'99 contest, the classifier learning contest, see here.

The purpose of the knowledge discovery contest held in conjunction with the KDD'99 conference was to showcase methods for discovering higher-level knowledge from data. Participants were asked to apply a range of knowledge discovery techniques to the same data used in the 1998 competition. They were free to exploit background knowledge about the domain.  For more information, see the detailed knowledge discovery contest description.  (Note added in May 2000: the link to the left is broken, but the dataset is still available here.)

The deliverable that each participating team submitted was a report explaining the team's results.  Three reports were submitted, from three companies.

The contest committee named two submissions as co-winners.  In no particular order:

•  "KDD'99 Competition: Knowledge Discovery Contest" by Jim Georges and Anne H. Milley of SAS Institute, Inc.  (This paper is available directly.)
•  "Knowledge Discovery in a Charitable Organization's Donor Database" by Saharon Rosset and Aron Inger of Amdocs (Israel) Ltd.

• "Profiling Your Customers Using Bayesian Networks" by Paola Sebastiani, Marco Ramoni, and Alexander Crea of Bayesware Ltd.