Features

Following the very successful contest in 2004, UCSD is again holding a data mining competition sponsored by Fair Isaac, Inc. The contest is open to everyone. University of California students and affiliates can win cash prizes.

There are two predictive modeling tasks, both involving real-world CRM data. Essentially, you are given data about accounts, and transactions for each account. The first task is to predict which accounts are "bad" (e.g. bankrupt, or the victim of identity theft, etc.) The second task is to predict precisely at what point in time each account goes bad (or that it never goes bad).

The deadline for submitting entries is June 30, 2005.

For more information, visit http://mill.ucsd.edu. You can track the success rate of the current leaders in real-time at http://mill.ucsd.edu/index.php?page=Leaderboard

Contact email is datacontest@cs.ucsd.edu