Constructing a Fuzzy Rule System from Examples

Authors:
Te-Min Chang;Yuehwern Yih
Affiliations:
School of Industrial Engineering, Purdue University, West Lafayette, IN 47907, USA;School of Industrial Engineering, Purdue University, West Lafayette, IN 47907, USA
Venue:
Integrated Computer-Aided Engineering
Year:
1999

Citing 4
Cited 0

Fuzzy Logic

Computer
Structure identification of fuzzy model

Fuzzy Sets and Systems
Incremental Induction of Decision Trees

Machine Learning
Induction of Decision Trees

Machine Learning

Quantified Score

Hi-index	0.02

Visualization

Abstract

A general framework to construct fuzzy rule systems automatically from given examples is proposed in this paper. The objective is to generate fuzzy systems with good mapping ability and generalization ability as well. The procedure consists of five steps. Cluster analysis of data is used for initial fuzzy region partitions. Resolution to rule conflict is included in the fuzzy defuzzification method. Inductive learning algorithm is incorporated to enhance generalization ability of fuzzy systems. System performance is iteratively improved by further partitioning fuzzy regions. Ineffective attributes can be implicated by the tree structure resulting from the learning algorithm and eliminated without deteriorating system performance. Several experiments are conducted to show advantages of this framework.