Multiobjective fuzzy linear regression analysis for fuzzy input-output data (Q1193642): Difference between revisions

The paper generalizes the so-called Tanaka-approach in fuzzy linear regression [see, e.g., \textit{H. Tanaka}, ibid. 24, 363-375 (1987; Zbl 0633.93060)] which consists, roughly speaking, in finding a fuzzy linear functional relationship with minimal possible spread, which covers all given fuzzy data at least to a given degree. Assuming \(L\)-shaped fuzzy numbers for the interesting fuzzy quantities, this leads to linear programming problems. Whereas Tanaka has considered models with crisp input (independent variable) and fuzzy output (dependent variable) the present paper deals with fuzzy input and fuzzy output, using three indices for the equality between fuzzy numbers based on possibility or necessity. A numerical example is given.