M.Sc. Tezi Görüntüleme

      In simple and multiple linear regression analysis, if conventional methods are using, parameter estimation for linear regression model is based on some assumptions. In the process of the estimation unknown parameters for linear regression models, if the data have different structure from normal distribution, it is requires to go beyond the classical resolution in the estimation process. In addition, the existence of contradictory observations in the data set in crease the importance of the method used in the process. In such cases, analysis methods based on fuzzy logic are known as alternative methods. In this study, in the case of any of independent variables have Pareto distribution and there were outlier observations in data set an algorithm has been suggested to define the unknown parameters of multiple linear regression model. The estimates obtained from this algorithm are compared with estimates obtained from existing methods.