Comparison of classifier methods: a case study in handwritten digit recognition
Léon Bottou, Corinna Cortes(AT&T (United States)), J. S. Denker(AT&T (United States)), Harris Drucker(AT&T (United States)), Isabelle Guyon(AT&T (United States)), L. D. Jackel(AT&T (United States)), Yann LeCun(AT&T (United States)), Urs Müller(AT&T (United States)), Eduard Säckinger(AT&T (United States)), P. Simard(École Polytechnique Fédérale de Lausanne), Vladimir Vapnik(École Polytechnique Fédérale de Lausanne)
Cited by 626
Abstract
This paper compares the performance of several classifier algorithms on a standard database of handwritten digits. We consider not only raw accuracy, but also training time, recognition time, and memory requirements. When available, we report measurements of the fraction of patterns that must be rejected so that the remaining patterns have misclassification rates less than a given threshold.