Fourier analysis of the generalized CMAC neural network

被引：11

作者：

Gonzalez-Serrano, FJ

Figueiras-Vidal, AR

Artes-Rodriguez, A

机构：

[1] UPM, DSSR ETSI Telecomunicac, Madrid, Spain

[2] Univ Vigo, DTC ETSI Telecomunicac, Vigo, Spain

[3] Univ Carlos III Madrid, EPS, ATSC DI, Leganes, Spain

来源：

NEURAL NETWORKS | 1998年 / 11卷 / 03期

关键词：

CMAC; associative memory; neural networks; function approximation; modeling capabilities;

D O I：

10.1016/S0893-6080(98)00017-3

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 [模式识别与智能系统]; 0812 [计算机科学与技术]; 0835 [软件工程]; 1405 [智能科学与技术];

摘要：

The Cerebellar Model Articulation Controller (CMAC) is a simple and fast neural network: these characteristics have extended its successful applications, while the analysis of its representation capabilities, as for many other neural networks, did not follow a similar development. In this article we discover the close parallelism between the representation of a function by a Generalized CMAC (GCMAC) and Nyquist sampling theory: discussing the role of different parameters and components of the network according to this similarity. The consideration of a representative example shows how the parallelism can be used to design a GCMAC;: adapted to its particular application. (C) 1998 Elsevier Science Ltd. All rights reserved.

引用

页码：391 / 396

页数：6

共 18 条

[1]

Albus J. S., 1975, Transactions of the ASME. Series G, Journal of Dynamic Systems, Measurement and Control, V97, P220, DOI 10.1115/1.3426922

[2]

MECHANISMS OF PLANNING AND PROBLEM-SOLVING IN THE BRAIN [J].