Determination of used car sale prices by fuzzy linear regression analysis

Linear regression is an effective statistical method that allows us to produce predictive values for the dependent variable by analysing the relationship between the dependent variable and one or more independent variables. The effectiveness of the linear regression method depends on the necessary assumptions. However, in many areas of daily life, it is very difficult to find data sets that meet these assumptions. Fuzzy linear regression was found as a result of applying L.A. Zadeh's fuzzy set theory to linear regression methods. Fuzzy linear regression is a method that can create a model without the need for assumptions required for linear regression in cases where the exact relationships between variables cannot be defined. In the theoretical part of this study, linear regression, fuzzy logic and its applications, fuzzy linear regression models are analysed. In the application part, second-hand vehicle sales price determination was performed using an open access data set. Linear regression assumptions were first applied to the data set and then a linear regression model was obtained. Secondly, fuzzy linear regression analysis was performed by applying different fuzziness criteria to the same data set. The models were evaluated by finding the sum of error squares