Resolution of fuzzy regression model

Abstract Fuzzy linear regression was originally introduced by Tanaka et al. To cope with different types of input–output information, several approaches to fuzzy linear regression have been proposed. In this paper, a type of problem with crisp input and fuzzy output described by Tanaka is considered of which a modified fuzzy least square method was proposed for solution. It shows that with such an approach the predictability in the new model is better than Tanaka’s and its computation efficiency is better than the conventional fuzzy least square method.