DocumentCode :
3069808
Title :
How to get an optimal neural network for a given application
Author :
Smeliansky, Ruslan L. ; Umnov, Nickolay A.
Author_Institution :
Dept. of Comput. Sci., Moscow State Univ., Russia
fYear :
1995
fDate :
20-23 Sep 1995
Firstpage :
329
Lastpage :
336
Abstract :
The paper presents a problem of optimal neural network (NN) design for a given application. The application is described as a mapping RN→R. The design process consists of two stages: selection of NN models which are able to approximate the given application, and an optimisation criterion estimation for each selected model with the further choosing of the best NN. The structure of an instrumental system for optimal NN design is proposed
Keywords :
graph theory; learning (artificial intelligence); minimisation; neural nets; design process; optimal neural network; optimisation criterion estimation; Application software; Cellular neural networks; Computer networks; Instruments; Neural networks; Neurons; Process design; State-space methods; Testing; Transfer functions;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Neuroinformatics and Neurocomputers, 1995., Second International Symposium on
Conference_Location :
Rostov on Don
Print_ISBN :
0-7803-2512-5
Type :
conf
DOI :
10.1109/ISNINC.1995.480877
Filename :
480877
Link To Document :
https://search.ricest.ac.ir/dl/search/defaultta.aspx?DTC=49&DC=3069808
بازگشت