Fuzzy linear regression was originally introduced by Tanaka, Uejima. a
nd Asai [IEEE Trans. Systems Man. Cybern. 12 (1982) 903 907]. In subse
quent years, several different approaches to fuzzy linear regression h
ave been proposed. The purpose of this paper is to review and examine
some of these formulations, to discuss their strengths and weaknesses
relative to each other, and to suggest possible improvements. In addit
ion, we compare and contrast these methods to the method of ordinary l
east squares regression.