Learning fuzzy classification rules from labeled data

J. Roubos, Magne Setnes, J. Abonyi

Research output: Contribution to journalArticle

134 Citations (Scopus)

Abstract

The automatic design of fuzzy rule-based classification systems based on labeled data is considered. It is recognized that both classification performance and interpretability are of major importance and effort is made to keep the resulting rule bases small and comprehensible. For this purpose, an iterative approach for developing fuzzy classifiers is proposed. The initial model is derived from the data and subsequently, feature selection and rule-base simplification are applied to reduce the model, while a genetic algorithm is used for parameter optimization. An application to the Wine data classification problem is shown.

Original languageEnglish
Pages (from-to)77-93
Number of pages17
JournalInformation Sciences
Volume150
Issue number1-2
DOIs
Publication statusPublished - Mar 2003

Fingerprint

Fuzzy Classification
Rule Base
Classification Rules
Fuzzy Rules
Fuzzy Classifier
Data Classification
Interpretability
Parameter Optimization
Classification Problems
Feature Selection
Simplification
Wine
Genetic Algorithm
Fuzzy rules
Feature extraction
Classifiers
Genetic algorithms
Model
Learning
Design

Keywords

  • Compact fuzzy classifier
  • Genetic algorithm
  • Linguistic model
  • Similarity-driven rule-base reduction
  • Wine data

ASJC Scopus subject areas

  • Statistics and Probability
  • Electrical and Electronic Engineering
  • Statistics, Probability and Uncertainty
  • Information Systems and Management
  • Information Systems
  • Computer Science Applications
  • Artificial Intelligence

Cite this

Learning fuzzy classification rules from labeled data. / Roubos, J.; Setnes, Magne; Abonyi, J.

In: Information Sciences, Vol. 150, No. 1-2, 03.2003, p. 77-93.

Research output: Contribution to journalArticle

@article{45ea50e592154144a3ed3fa99d631b1f,
title = "Learning fuzzy classification rules from labeled data",
abstract = "The automatic design of fuzzy rule-based classification systems based on labeled data is considered. It is recognized that both classification performance and interpretability are of major importance and effort is made to keep the resulting rule bases small and comprehensible. For this purpose, an iterative approach for developing fuzzy classifiers is proposed. The initial model is derived from the data and subsequently, feature selection and rule-base simplification are applied to reduce the model, while a genetic algorithm is used for parameter optimization. An application to the Wine data classification problem is shown.",
keywords = "Compact fuzzy classifier, Genetic algorithm, Linguistic model, Similarity-driven rule-base reduction, Wine data",
author = "J. Roubos and Magne Setnes and J. Abonyi",
year = "2003",
month = "3",
doi = "10.1016/S0020-0255(02)00369-9",
language = "English",
volume = "150",
pages = "77--93",
journal = "Information Sciences",
issn = "0020-0255",
publisher = "Elsevier Inc.",
number = "1-2",

}

TY - JOUR

T1 - Learning fuzzy classification rules from labeled data

AU - Roubos, J.

AU - Setnes, Magne

AU - Abonyi, J.

PY - 2003/3

Y1 - 2003/3

N2 - The automatic design of fuzzy rule-based classification systems based on labeled data is considered. It is recognized that both classification performance and interpretability are of major importance and effort is made to keep the resulting rule bases small and comprehensible. For this purpose, an iterative approach for developing fuzzy classifiers is proposed. The initial model is derived from the data and subsequently, feature selection and rule-base simplification are applied to reduce the model, while a genetic algorithm is used for parameter optimization. An application to the Wine data classification problem is shown.

AB - The automatic design of fuzzy rule-based classification systems based on labeled data is considered. It is recognized that both classification performance and interpretability are of major importance and effort is made to keep the resulting rule bases small and comprehensible. For this purpose, an iterative approach for developing fuzzy classifiers is proposed. The initial model is derived from the data and subsequently, feature selection and rule-base simplification are applied to reduce the model, while a genetic algorithm is used for parameter optimization. An application to the Wine data classification problem is shown.

KW - Compact fuzzy classifier

KW - Genetic algorithm

KW - Linguistic model

KW - Similarity-driven rule-base reduction

KW - Wine data

UR - http://www.scopus.com/inward/record.url?scp=0037361892&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0037361892&partnerID=8YFLogxK

U2 - 10.1016/S0020-0255(02)00369-9

DO - 10.1016/S0020-0255(02)00369-9

M3 - Article

AN - SCOPUS:0037361892

VL - 150

SP - 77

EP - 93

JO - Information Sciences

JF - Information Sciences

SN - 0020-0255

IS - 1-2

ER -