Artificial neural networks in problems of material objects implementation. Part 1. Networking principles and Classification

Abstract

The main historical development stages of artificial neural networks are presented. The mathematical model of artificial neurons and the basic elements of neural networks are considered. Typical activation functions are characterized with their advantages and disadvantages. One of possible problem solution of the neural networks classification is given according to different classification features. Expediency of neural networks using in problems of material objects implementation is substantiated. That will allow in the nearest future to bring engineering intuition and experience components in the calculations.

Authors and Affiliations

V. Adamenko, G. Mirskikh

Keywords

Related Articles

Evaluation of sensibility and resolution of the control for magnetic protection of documents by the magneto-optic television system

Introduction. The objective evaluation of magneto-optical television system technical characteristics was needed in the development. Novelty of the system prevents the use of standard techniques for this. The task of thi...

Modern devices of optimum filtration for the active radar system

The principle of construction the matched filter and correlator, for the active radar system operating with a broadband noise signal is esteemed. The example of construction a сhan-nel of processing on the basis of micro...

Information loss in radial velocity canal

The radial velocity canal work in conditions of information loss is considered and analyzed. Information velocity loss about object’s radial velocity is defined.

Complex Discrete Normal Ortogonal Transform

The algorithm of forming matrix operator of discrete normal orthogonal complex transform is offered, possibility of its application for the authentication of person after the dynamically entered signature is shown.

Peculiarities of signal formation of the autodyne short-range radar with linear frequency modulation

Research results of signal formation peculiarities in autodyne short-range radars with linear frequency modulation by non-symmetric saw-tooth law are presented. Calculation expressions for autodyne signals are obtained i...

Download PDF file
  • EP ID EP309323
  • DOI 10.20535/RADAP.2011.47.176-189
  • Views 48
  • Downloads 0

How To Cite

V. Adamenko, G. Mirskikh (2011). Artificial neural networks in problems of material objects implementation. Part 1. Networking principles and Classification. Вісник НТУУ КПІ. Серія Радіотехніка, Радіоапаратобудування, 0(47), 176-189. https://europub.co.uk/articles/-A-309323