The Comparison and an Application of the Classic Least Squares Method and the Fuzzy Least Squares Method

6 th International Researchers, Statisticians and Young Statisticians Congress (IRSYSC2022), Antalya, Türkiye, 3 - 06 Kasım 2022, ss.86

One of the numerical methods that we use in classical Linear Regression analysis when

measuring the relationship between quantitative variables is the Least Squares Method. The main

purpose of this method is to provide coefficient estimation values that help to reduce error terms.

Since it will be difficult for us to reach a conclusion with the classical regression model when

qualitative data are in question, the Fuzzy Regression Model comes across as an alternative way. The

Fuzzy Regression Model emerged as a result of converting the concept of Fuzzy Logic into regression

analysis. For example; In fuzzy regression analysis, the fact that the data do not have a normal

distribution does not prevent the application of this analysis. In the analysis of the fuzzy regression

model, the regression coefficients and the estimated values are also fuzzy numbers. The most

common method used to implement this analysis is the Fuzzy Least Squares Method developed by

Diamond (1988). In this method, it tries to decode the fuzzy gap between the estimated fuzzy

independent variable value and the observed values in such a way that it is the smallest.