The linear separability problem: some testing methods

Date

2006-03-01

Advisors

Journal Title

Journal ISSN

ISSN

1045-9227

Volume Title

Publisher

IEEE

Type

Article

Peer reviewed

Yes

Abstract

Description

This article presents an analysis of some of the methods for testing linear separability. A single layer perceptron neural network can be used for creating a classification model when the classes at hand are linearly separable. Since the RDP neural network is based on linearly separable subsets within a non linearly separable set, the performance of the method used for searching these subsets is of great importance in order to minimise convergence time, and maximise the level of generalisation. It appears in one of the leading journals of Neural Networks with an impact factor of 2.620.

Keywords

RAE 2008, UoA 23 Computer Science and Informatics, class of separability, computational geometry, Fisher linear discriminant, linear programming

Citation

Elizondo, D.A. (2006) The linear separability problem: some testing methods. IEEE Transactions on Neural Networks, 17(2), pp. 330-344.

Rights

Research Institute