NN-driven fuzzy reasoning

Hideyuki Takagi, Isao Hayashi

Research output: Contribution to journalArticle

387 Citations (Scopus)

Abstract

A new fuzzy reasoning that can solve two problems of conventional fuzzy reasoning by combining an artificial neural network (NN) and fuzzy reasoning is proposed. These problems are (1) the lack of design for a membership function except a heuristic approach and (2) the lack of adaptability for possible changes in the reasoning environment. The proposed fuzzy reasoning approach solves these problems by using the learning function and nonlinearity of an NN. First, the problems involved in conventional fuzzy reasoning and the NN used in this paper are identified. Then a proposed algorithm is formulated and a concrete explanation using realistic data is developed. An example structure of an NN-driven fuzzy reasoning system is given, and two applications of this method are presented. This new fuzzy reasoning is capable of automatic determination of inference rules and adjustment according to the time-variant reasoning environment because of the use of NN in fuzzy reasoning. This proposed method can be applied to NN modeling and AI and is considered from the standpoint of the explicit incorporation of knowledge into the NN structure.

Original languageEnglish
Pages (from-to)191-212
Number of pages22
JournalInternational Journal of Approximate Reasoning
Volume5
Issue number3
DOIs
Publication statusPublished - Jan 1 1991
Externally publishedYes

Fingerprint

Fuzzy Reasoning
Fuzzy neural networks
Neural Networks
Neural networks
Reasoning
Membership functions
Inference Rules
Network Modeling
Adaptability
Membership Function
Network Structure
Concretes
Artificial Neural Network
Adjustment
Nonlinearity
Heuristics

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Software
  • Artificial Intelligence
  • Applied Mathematics

Cite this

NN-driven fuzzy reasoning. / Takagi, Hideyuki; Hayashi, Isao.

In: International Journal of Approximate Reasoning, Vol. 5, No. 3, 01.01.1991, p. 191-212.

Research output: Contribution to journalArticle

Takagi, Hideyuki ; Hayashi, Isao. / NN-driven fuzzy reasoning. In: International Journal of Approximate Reasoning. 1991 ; Vol. 5, No. 3. pp. 191-212.
@article{c2c938e9ceb849c2abc087fd20ebbdfe,
title = "NN-driven fuzzy reasoning",
abstract = "A new fuzzy reasoning that can solve two problems of conventional fuzzy reasoning by combining an artificial neural network (NN) and fuzzy reasoning is proposed. These problems are (1) the lack of design for a membership function except a heuristic approach and (2) the lack of adaptability for possible changes in the reasoning environment. The proposed fuzzy reasoning approach solves these problems by using the learning function and nonlinearity of an NN. First, the problems involved in conventional fuzzy reasoning and the NN used in this paper are identified. Then a proposed algorithm is formulated and a concrete explanation using realistic data is developed. An example structure of an NN-driven fuzzy reasoning system is given, and two applications of this method are presented. This new fuzzy reasoning is capable of automatic determination of inference rules and adjustment according to the time-variant reasoning environment because of the use of NN in fuzzy reasoning. This proposed method can be applied to NN modeling and AI and is considered from the standpoint of the explicit incorporation of knowledge into the NN structure.",
author = "Hideyuki Takagi and Isao Hayashi",
year = "1991",
month = "1",
day = "1",
doi = "10.1016/0888-613X(91)90008-A",
language = "English",
volume = "5",
pages = "191--212",
journal = "International Journal of Approximate Reasoning",
issn = "0888-613X",
publisher = "Elsevier Inc.",
number = "3",

}

TY - JOUR

T1 - NN-driven fuzzy reasoning

AU - Takagi, Hideyuki

AU - Hayashi, Isao

PY - 1991/1/1

Y1 - 1991/1/1

N2 - A new fuzzy reasoning that can solve two problems of conventional fuzzy reasoning by combining an artificial neural network (NN) and fuzzy reasoning is proposed. These problems are (1) the lack of design for a membership function except a heuristic approach and (2) the lack of adaptability for possible changes in the reasoning environment. The proposed fuzzy reasoning approach solves these problems by using the learning function and nonlinearity of an NN. First, the problems involved in conventional fuzzy reasoning and the NN used in this paper are identified. Then a proposed algorithm is formulated and a concrete explanation using realistic data is developed. An example structure of an NN-driven fuzzy reasoning system is given, and two applications of this method are presented. This new fuzzy reasoning is capable of automatic determination of inference rules and adjustment according to the time-variant reasoning environment because of the use of NN in fuzzy reasoning. This proposed method can be applied to NN modeling and AI and is considered from the standpoint of the explicit incorporation of knowledge into the NN structure.

AB - A new fuzzy reasoning that can solve two problems of conventional fuzzy reasoning by combining an artificial neural network (NN) and fuzzy reasoning is proposed. These problems are (1) the lack of design for a membership function except a heuristic approach and (2) the lack of adaptability for possible changes in the reasoning environment. The proposed fuzzy reasoning approach solves these problems by using the learning function and nonlinearity of an NN. First, the problems involved in conventional fuzzy reasoning and the NN used in this paper are identified. Then a proposed algorithm is formulated and a concrete explanation using realistic data is developed. An example structure of an NN-driven fuzzy reasoning system is given, and two applications of this method are presented. This new fuzzy reasoning is capable of automatic determination of inference rules and adjustment according to the time-variant reasoning environment because of the use of NN in fuzzy reasoning. This proposed method can be applied to NN modeling and AI and is considered from the standpoint of the explicit incorporation of knowledge into the NN structure.

UR - http://www.scopus.com/inward/record.url?scp=44949275618&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=44949275618&partnerID=8YFLogxK

U2 - 10.1016/0888-613X(91)90008-A

DO - 10.1016/0888-613X(91)90008-A

M3 - Article

VL - 5

SP - 191

EP - 212

JO - International Journal of Approximate Reasoning

JF - International Journal of Approximate Reasoning

SN - 0888-613X

IS - 3

ER -