Akademik Veri Yönetim Sistemi
XXIII. IBANESS Congress Series on Economics, Business and Management , 15 - 16 Mart 2025, ss.783-786, (Özet Bildiri)
This study has two main purposes; the first is to analyze the impacts of different fuzziness levels on parameter estimates in Fuzzy Regression Analysis, and the second is to show that Fuzzy Regression Analysis can produce more accurate predictions than Least Squares Regression Method when there is not enough data. The application was carried out with real data through a model taken from the literature. The dependent variable of the model is; GDP (representing economic growth), and the independent variables are; education expenditures, inflation rate and the unemployment rate. These are the variables of the model included in the work of Öztürk, Kalaycı and Korkmaz published in 2017. Although there is no problem in accessing data regarding these variables, it was studied with a particularly small number of data to demonstrate the superiority of Fuzzy Regression Analysis over the Least Squares Regression Method. As it is known, when there is not enough data, the Least Squares Regression Method suffers from assumption distortions. The data are annual data for the period 2012-2022 and were obtained from the World Bank. RStudio and LINGO were used for the predictions of the models.