ARTIFICIAL NEURAL NETWORKS: INTRODUCTION AND USE OF PERCEPTRON ALGORITHM IN BIOSSYSTEMS
DOI:
https://doi.org/10.18011/bioeng2010v4n2p87-95Keywords:
Artificial neuron, Classification of groups, Learning, OperationAbstract
Artificial Neural Networks are computational model inspired by way living organisms manipulate the information received, and with that, presents learning capability, adaptability and generalization of knowledge. Moreover, the operation of an Artificial Neural Network is based on parallelism, resembling as the brain deals with information received by the neurons. The present work describes the study of artificial neural network Perceptron as a linear classifier to distinguish two types of fruits (orange (citrus sinensis Osbeck) and tangerine 'Ponkan' (citrus reticulate Blanco)) using two basic processes performed by an artificial neural network, the phases of training or learning and operation phase. The main objective is to show the functioning of these phases.
Downloads
References
Haykin S., Neural Networks: a comprehensive foundation. New York: MacMillan. College Publishing Co., 1999.
BRAGA, A. de P.; CARVALHO, A. P. de L. F. de; LUDERMIR, T. B. Redes Neurais Artificiais – Teoria e Aplicações. 2.ed. Rio de Janeiro: LTC, 2007. p. 1-20
Mcculloch W. S. e Pitts, W. A. Logical Calculus of the Ideas Immanent in Nervous Activity, Bulletin of Mathematical Biophysics, 1943, No. 9, pp. 127-147.
Minussi C. R. Redes Neurais: Introdução e principais conceitos. Faculdade de Engenharia de Ilha Solteira. Notas de aula – UNESP. 2003.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2010 The Authors
This work is licensed under a Creative Commons Attribution 4.0 International License.
By publishing in this journal, authors agree to the following terms:
a) Authors retain copyright and grant the journal the right of first publication. The work is simultaneously licensed under the Creative Commons Attribution License, which permits sharing and adaptation of the work with appropriate credit to the authors and the journal.
b) Authors may enter into separate, additional agreements for non-exclusive distribution of the published version of the work (e.g., posting to an institutional repository or inclusion in a book), provided that proper credit is given to the original publication in this journal.
