Small-sample error estimation: mythology versus mathematics Conference Paper uri icon

abstract

  • Error estimation is a key aspect of statistical pattern recognition. The true classification error rate is usually unavailable since it depends on the unknown feature-label distribution. Hence, one needs to estimate the error rate from the available sample data. This paper presents a concise, mathematically rigorous review of the subject of error estimation in statistical pattern recognition, pointing to the pitfalls that arise in small-sample settings due to the use of "rules of thumb" and a neglect for proper mathematical understanding of the problem.

name of conference

  • Mathematical Methods in Pattern and Image Analysis

published proceedings

  • Proceedings of SPIE

author list (cited authors)

  • Braga-Neto, U.

citation count

  • 5

complete list of authors

  • Braga-Neto, Ulisses

editor list (cited editors)

  • Astola, J. T., Tabus, I., & Barrera, J.

publication date

  • January 2005