|
CMU-CS-01-102
Computer Science Department
School of Computer Science, Carnegie Mellon University
CMU-CS-01-102
Bounds for Averaging Classifiers
John Langford+, Matthias Seeger*
January 2001
CMU-CS-01-102.ps
CMU-CS-01-102.pdf
Keywords: PAC bound, Maximum entropy discrimination, averaging
hypotheses
We present a generalized PAC bound for averaging classifiers which
applies to base hypotheses with a bounded real valued output. In
addition, we discuss several methods for quantitatively tightening the
bound. In the process, a tightened version of the PAC-Bayes bound
is proved.
15 pages
+jcl@cs.cmu.edu, www.cs.cmu.edu/~jcl
*Institute for Adaptive and Neural Computation, University of Edinburgh,
seeger@dai.ed.ac.uk, www.dai.ed.ac.uk/homes/seeger/
|