A NOTE ON EVALUATION OF FUZZY LINEAR REGRESSION MODELS BY COMPARING MEMBERSHIP FUNCTIONS

نویسندگان
Hassanpour, H. ID 1
Malek, H. R. ID 2
Yaghoobi, M. A. ID 3

1 Department of Mathematics, University of Birjand, Birjand, Iran

2 Faculty of Basic Sciences, Shiraz University of Technology, Shiraz, Iran

3 Department of Statistics, Shahid Bahonar University of Kerman, Kerman, Iran

doi
10.22111/ijfs.2009.203
چکیده

Kim and Bishu (Fuzzy Sets and Systems 100 (1998) 343-352) proposeda modification of fuzzy linear regression analysis. Their modificationis based on a criterion of minimizing the difference of the fuzzy membershipvalues between the observed and estimated fuzzy numbers. We show that theirmethod often does not find acceptable fuzzy linear regression coefficients andto overcome this shortcoming, propose a modification. Finally, we present twonumerical examples to illustrate efficiency of the modified method.

کلیدواژه‌ها
Fuzzy linear regression, Fuzzy number, Least-squares method. This paper is supported in part by Fuzzy Systems and Applications Center of Excellence, Shahid Bahonar University of Kerman, Kerman, I.R. of Iran