Multivariate non-linear fuzzy regression: An evolutionary algorithm approach

被引：15

作者：

Buckley, JJ ^{[1
]}

Feuring, T

Hayashi, Y

机构：

[1] Univ Alabama, Dept Math, Birmingham, AL 35294 USA

[2] Univ Gesamthsch Siegen, Dept Elect Engn & Comp Sci, D-57068 Siegen, Germany

[3] Meiji Univ, Dept Comp Sci, Tama Ku, Kawasaki, Kanagawa 2148571, Japan

来源：

INTERNATIONAL JOURNAL OF UNCERTAINTY FUZZINESS AND KNOWLEDGE-BASED SYSTEMS | 1999年 / 7卷 / 02期

关键词：

fuzzy regression; multivariate regression; evolutionary algorithm;

D O I：

10.1142/S0218488599000076

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

We first argue that a very important class of fuzzy functions in multivariate non-linear fuzzy regression is the multivariate fuzzy polynomials. Given some data, generated by a multivariate fuzzy function, our evolutionary algorithm searches our library of multivariate fuzzy polynomials for the one that best fits this data. Tests of our multivariate non-linear fuzzy regression package are given when the multivariate fuzzy function is a multivariate fuzzy polynomial and all the fuzzy numbers are non-negative.

引用

页码：83 / 98

页数：16