Explicit overall risk minimization transductive bound
by
Sergio Decherchi
Dept Biophysical and Electronic Engineering, University of Genoa
Coauthors: Paolo Gastaldo, Sandro Ridella, Rodolfo Zunino

In recent years, approaches alternatives to full induction have reached an always increasing attention from the machine learning research community. Beside inductive learning schemes, exist the so called transductive learning: in this environment is not required generalization for every possible input, instead only achieving the best possible performance on a particular and known test data. In this work adapting the machinery of Vapnik theory and an upper bound on the hypergeometric distribution, an explicit formula will be obtained for overall risk minimization trasductive error.

Date received: April 24, 2008

Copyright © 2008 by the author(s). The author(s) of this document and the organizers of the conference have granted their consent to include this abstract in Atlas Conferences Inc. Document # caxc-71.