Interpolative reasoning approach to sparse general type-2 fuzzy rules based on the reduced grid representation

Author

Long Thanh Ngo ; Minh Ngoc Vu ; Hirota, Kaoru

Author_Institution

Dept. of Inf. Syst., Le Quy Don Tech. Univ., Hanoi, Vietnam

fYear

2013

fDate

15-18 Dec. 2013

Firstpage

72

Lastpage

78

Abstract

Interpolative reasoning is one of the most interested problems with various approaches for type-1 fuzzy sets, interval type-2 fuzzy sets, recently. However, the related methods have not mentioned general type-2 fuzzy sets yet because of their computational complexity. The paper deals with an approach to representation theorem of general type-2 fuzzy sets using the reduced grid. A computational schema for interpolative reasoning of sparse general type-2 fuzzy rules is also introduced. This schema is not depended on the shape of membership functions. Beside, the parallelizing schema for GPU platform is proposed to speedup the algorithms. The proposed methods are implemented on both of GPU and CPU platforms with various membership functions.

Keywords

computational complexity; fuzzy set theory; grid computing; inference mechanisms; interpolation; parallel processing; CPU platforms; GPU platform; computational complexity; computational schema; general type-2 fuzzy sets; interpolative reasoning; interval type-2 fuzzy sets; membership functions; parallel computing; parallelizing schema; reduced grid representation; sparse general type-2 fuzzy rules; type-1 fuzzy sets; Cognition; Computational complexity; Fuzzy sets; Graphics processing units; Indexes; Interpolation; Pattern recognition; GPU parallel computing; Interpolative reasoning; general type-2 fuzzy sets; sparse fuzzy rule;

fLanguage

English

Publisher

ieee

Conference_Titel

Soft Computing and Pattern Recognition (SoCPaR), 2013 International Conference of

Conference_Location

Hanoi

Print_ISBN

978-1-4799-3399-0

Type

conf

DOI

10.1109/SOCPAR.2013.7054104

Filename

7054104