Fuzzy number neural networks

被引：15

作者：

Dunyak, J ^{[1
]}

Wunsch, D

机构：

[1] Texas Tech Univ, Dept Math, Lubbock, TX 79409 USA

[2] Texas Tech Univ, Dept Elect Engn, Lubbock, TX 79409 USA

来源：

FUZZY SETS AND SYSTEMS | 1999年 / 108卷 / 01期

关键词：

fuzzy numbers; neural networks; learning;

D O I：

10.1016/S0165-0114(97)00339-4

中图分类号：

TP301 [理论、方法];

学科分类号：

081202 ;

摘要：

This paper presents a practical algorithm for training neural networks with fuzzy number weights, inputs, and outputs. Typically, fuzzy number neural networks are difficult to train because of the many a-cut constraints implied by the fuzzy weights. A transformation is used to eliminate these constraints and allow use of standard unconstrained optimization methods. The algorithm is demonstrated on a three-layer network. (C) 1999 Elsevier Science B.V. All rights reserved.

引用

页码：49 / 58

页数：10

共 10 条

[1] FUZZY NEURAL NETWORKS - A SURVEY [J].

BUCKLEY, JJ ;

HAYASHI, Y .

FUZZY SETS AND SYSTEMS, 1994, 66 (01) :1-13

[2]

BUCKLEY JJ, UNPUB FUZZY BACK PRO

[3]

FEURING T, 1996, P ICNN 96, V2, P1061

[4] FUZZY NEURAL NETWORK WITH FUZZY SIGNALS AND WEIGHTS [J].

HAYASHI, Y ;

BUCKLEY, JJ ;

CZOGALA, E .

INTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS, 1993, 8 (04) :527-537

[5]

HAYASHI Y, 1992, P 2 INT C FUZZ LOG N, P73

[6]

HAYASHI Y, 1992, P INT JOINT C NEURAL, V2, P696

[7]

ISHIBUCHI H, 1993, P IEEE INT C NEUR NE, V1, P1650

[8]

ISHIBUCHI H, 1993, 5TH P IFSA WORLD C S, V1, P147

[9]

ISHIBUCHI H, 1996, P INT C NEUR NETW, V2, P1191

[10]

WERBOS P, 1990, P IEEE, V78

← 1 →