A New Approach to Fuzzy Regression Models with Application to Business Cycle Analysis

Abstract

Recently, fuzzy regression analysis has been largely applied in the modeling of economic or financial data. However, those data often exhibit certain kinds of linguistic terms, for instance: very good, a little reclining or stable, in the business cycle or the growth rate of GDP, etc. The goal of this paper is to construct a fuzzy regression model by fuzzy parameters estimation using the fuzzy samples. It deals with imprecise measurement of observed variables, fuzzy least square estimation and nonparametric methods. This is different from the assumptions as well as the estimation techniques of the classical analysis. Empirical results demonstrate that our new approach is efficient and more realistic than the traditional regression analysis.