April 25, 2012 -
9:30 am to 11:00 am
Location: Cupples I, Room 6 |
Host: Prof. Nan Lin
Abstract: We formulate the ranking problem in a rigorous statistical framework. The goal is to learn a ranking rule to have minimum ranking risk. The natural estimates of the risk have the form of U-statistics. With results of the theory of U-processes, we investigate the consistency of empirical risk minimizers. A fast rate of convergence can be achieved under some noise assumptions. Finally, we investigate the consistency of the augmented ranking risk minimizer, which can reduce the computational complexity.
Contact Us | Home Page
Department of Mathematics | Washington University in St. Louis | One Brookings Drive, St. Louis, MO 63130-4899 |
contact@math.wustl.edu