Domain based classification

Abstract

The majority of traditional classification ru les minimizing the expected probability of error (0-1 loss) are inappropriate if the class probability distributions are ill-defined or impossible to estimate. We argue that in such cases class domains should be used instead of class distributions or densities to construct a reliable decision function. Proposals are presented for some evaluation criteria and classifier learning schemes, illustrated by an example.

arXiv comment8 pages, unpublished paper written in 2005, discussing a significant, still almost not studied problem. missing reference links corrected