TITLE: Design and Evaluation of Neural Classifiers

AUTHORS: M. Hintz-Madsen, M.W. Pedersen, L.K. Hansen and J. Larsen
         Department of Mathematical Modelling, Building 349
         Technical University of Denmark
         DK-2800 Lyngby, Denmark
         emails: hintz@ei.dtu.dk, with@ei.dtu.dk, 
                 lkhansen@ei.dtu.dk, jl@imm.dtu.dk


ABSTRACT:
In this paper we propose a method for design of feed-forward neural 
classifiers based on regularization and adaptive architectures.
Using a penalized maximum likelihood scheme we derive a modified form
of the entropic error measure and an algebraic estimate of the
test error. In conjunction with Optimal Brain Damage pruning the test error
estimate is used to optimize the network architecture. The scheme is
evaluated on an artificial and a real world problem.

To appear in Proceedings of the IEEE Workshop on Neural Networks for Signal
Processing VI, Piscataway, New Jersey: IEEE, 1996.