Advances and Applications in Statistics

The Advances and Applications in Statistics is an internationally recognized journal indexed in the Emerging Sources Citation Index (ESCI). It provides a platform for original research papers and survey articles in all areas of statistics, both computational and experimental in nature.

Submit Article

SUFFICIENT CONDITIONS FOR ERGODICITY OF THE SINGLE-LAYER PERCEPTRON WEIGHT SEQUENCE IN INFINITE NON-SEPARABLE CLASSIFICATION PROBLEMS

Authors

  • Rieken S. Venema

DOI:

https://doi.org/10.17654/0972361723036

Abstract

In this paper, we study the asymptotic behavior of the weight sequence of a single-layer perceptron. If the perceptron is used for the classification of two infinite populations that cannot be linearly separated, the weights do not converge but we show that their probability distributions approach a steady state. Assuming that the input vectors form an i.i.d. sequence, we shall give sufficient conditions under which the perceptron weight process is ergodic.

References

F. Rosenblatt, Principles of Neurodynamics: Perceptrons and the Theory of Brain Mechanisms, Spartan Books, New York, NY, 1962.

M. Minsky and S. Papert, Perceptrons: An Introduction to Computational Geometry, The MIT Press, Cambridge, MA, 1969.

V. P. Roychowdhury, K.Y. Siu and T. Kailath, Classification of linearly nonseparable patterns by linear threshold elements, IEEE Transactions on Neural Networks 6(2) (1995), 318-331.

C. Cortes and V. Vapnik, Support-vector networks, Machine Learning 20(3) (1995), 273-297.

S. I. Gallant, Perceptron-based learning algorithms, IEEE Transactions on Neural Networks 1(2) (1990), 179-191.

Y. Freund and R. Schapire, Large margin classification using the perceptron algorithm, Machine Learning 37(3) (1999), 277-296.

J. Yang, R. Parekh and V. Honavar, Comparison of performance of variants of single-layer perceptron algorithms on non-separable data, Neural, Parallel, and Scientific Computation 8 (2000), 415-438.

J. J. Shynk and N. J. Bershad, Stationary points of a single-layer perceptron for nonseparable data models, Neural Networks 6(2) (1993), 189-202.

N. J. Bershad and J. J. Shynk, Performance analysis of a converged single-layer perceptron for nonseparable data models with bias terms, IEEE Transactions on Signal Processing 42(1) (1994), 175-188.

B. Efron, The Perceptron Correction Procedure in Non-separable Situations, Technical Report RADC-TDR-63-533, Rome Air Development Center, Rome NY, 1964.

R. M. Burton, H. G. Dehling and R. S. Venema, Perceptron algorithms for the classification of non-separable populations, Communications in Statistics – Stochastic Models 13(2) (1997), 205-222.

R. L. Tweedie, Sufficient conditions for ergodicity and recurrence of Markov chains on a general state space, Stochastic Processes and their Applications 3(4) (1975), 385-403.

R. L. Tweedie, R-theory for Markov chains on a general state space I: solidarity properties and R-recurrent chains, Annals of Probability 2 (1974), 840-864.

D. B. Pollard and R. L. Tweedie, R-theory for Markov chains on a topological state-space I, Journal of the London Mathematical Society 10 (1975), 389-400.

Downloads

Published

24-09-2025

Issue

Vol. 87 No. 2 (2023)

Section

Articles

License

Copyright (c) 2023 Pushpa Publishing House, Prayagraj, India

Creative Commons License

This work is licensed under a Creative Commons Attribution 4.0 International License.

____________________________

Licensing Terms:

Attribution: Credit Pushpa Publishing House as the original publisher, including title and author(s) if applicable.

No Derivatives: Modifying or creating derivative works not allowed without written permission.

Contact Pushpa Publishing House for more info or permissions.

How to Cite

SUFFICIENT CONDITIONS FOR ERGODICITY OF THE SINGLE-LAYER PERCEPTRON WEIGHT SEQUENCE IN INFINITE NON-SEPARABLE CLASSIFICATION PROBLEMS. (2025). Advances and Applications in Statistics , 87(2), 237-254. https://doi.org/10.17654/0972361723036