Structurization of the Covariance Matrix by Process Type and Block-Diagonal Models in the Classifier Design

Article type: Research Article

Authors: Saudargienė, Aušra

Affiliations: Institute of Mathematics and Informatics, Akademijos 4, 2600 Vilnius, Lithuania. E-mail: [email protected]

Abstract: Structurization of the sample covariance matrix reduces the number of the parameters to be estimated and, in a case the structurization assumptions are correct, improves small sample properties of a statistical linear classifier. Structured estimates of the sample covariance matrix are used to decorellate and scale the data, and to train a single layer perceptron classifier afterwards. In most from ten real world pattern classification problems tested, the structurization methodology applied together with the data transformations and subsequent use of the optimally stopped single layer perceptron resulted in a significant gain in comparison with the best statistical linear classifier – the regularized discriminant analysis.

Keywords: regularized discriminant analysis, single layer perceptron, generalization, covariance matrix, dimensionality, learning-set size

DOI: 10.3233/INF-1999-10208

Journal: Informatica, vol. 10, no. 2, pp. 245-269, 1999