On the Design of Interpretable Evolutionary Fuzzy Systems (I-EFS) with Improved Accuracy

Author

Shukla, P.K. ; Tripathi, S.P.

Author_Institution

Dept. of Inf. Technol., Babu Banarasi Das Northern India Inst. of Technol., Lucknow, India

fYear

2012

fDate

14-15 Sept. 2012

Firstpage

11

Lastpage

14

Abstract

Interpretability and accuracy are two important requirements during the development of fuzzy systems. This paper discusses various approaches related to the development of fuzzy systems in an Evolutionary Multiobjective Optimization (EMO) framework with good degree of interpretability and accuracy which are conflicting in their nature. This situation is well known as Interpretability-Accuracy (I-A) Trade-Off. Rule selection, rule learning, membership function tuning, fuzzy partition etc. are the major point of consideration to deal with this trade-off. Finally various recent issues related to this area in EMO framework are discussed.

Keywords

evolutionary computation; fuzzy reasoning; fuzzy set theory; fuzzy systems; knowledge based systems; optimisation; EMO framework; I-A trade-off; I-EF design; accuracy degree; evolutionary multiobjective optimization framework; interpretability degree; interpretability-accuracy trade-off; interpretable evolutionary fuzzy system design; Accuracy; Evolutionary computation; Fuzzy systems; Genetics; Optimization; Pragmatics; Tuning; Accuracy; Evolutionary Multi-objective Optimization (EMO); Inpterpretability; Interpretability-Accurcay (I-A) Trade-Off; Multi-Objective Evolutionary Algorithms (MOEA);

fLanguage

English

Publisher

ieee

Conference_Titel

Computing Sciences (ICCS), 2012 International Conference on

Conference_Location

Phagwara

Print_ISBN

978-1-4673-2647-6

Type

conf

DOI

10.1109/ICCS.2012.50

Filename

6391638