On training sample size and complexity of artificial neural net classifier

Article type: Research Article

Authors: Jain, Anil | Raudys, Šarūnas

Affiliations: Department of Computer Science, Michigan State University, East Lansing, MI 48824, USA | Department of Data Analysis, Institute of Mathematics and Informatics, 2600 Vilnius, Akademijos St.4, Lithuania

Abstract: Small training sample effects common in statistical classification and artificial neural network classifier design are discussed. A review of known small sample results are presented, and peaking phenomena related to the increase in the number of features and the number of neurons is discussed.

Keywords: feed forward neural nets, classification error training samples, dimensionality, complexity, peaning phenomenon

DOI: 10.3233/INF-1992-3301

Journal: Informatica, vol. 3, no. 3, pp. 301-337, 1992