Generalization Error of Randomized Linear Zero Empirical Error Classifier: Simple Asymptotics for Centered Data Case

Article type: Research Article

Authors: Dičiūnas, Valdas^; | Raudys, Šarūnas

Affiliations: Department of Computer Science, Vilnius University, Naugarduko 24, 2600 Vilnius, Lithuania. E-mail: valdas.diciunas@maf.vu.lt | Institute of Mathematics and Informatics, Akademijos 4, 2600 Vilnius, Lithuania. E-mail: raudys@das.mii.lt

Note: [] Part of this work was done while being at the Institute of Mathematics and Informatics, Vilnius, Lithuania.

Abstract: An estimation of the generalization performance of classifier is one of most important problems in pattern clasification and neural network training theory. In this paper we estimate the generalization error (mean expected probability of classification) for randomized linear zero empirical error (RLZEE) classifier which was considered by Raudys, Dičiūnas and Basalykas. Instead of “non-explicit” asymptotics of a generalization error of RLZEE classifier for centered multivariate spherically Gaussian classes proposed by Basalykas et al. (1996) we obtain an “explicit” and more simple asymptotics. We also present the numerical simulations illustrating our theoretical results and comparing them with each other and previously obtained results.

Keywords: randomized linear classifier, generalization error, Gaussian classes, probability of misclassification

DOI: 10.3233/INF-2000-11403

Journal: Informatica, vol. 11, no. 4, pp. 381-396, 2000