Robust regression estimators examining and application

Robust Regression Estimators Examining and Application It is known that in case the distribution assumption is being ensured, the method of least squares gives results representing the data sets. However, the least squares method which is sensitive to the outliers, may give misleading results when the normal distribution assumption is not provided. Therefore, in order to be away from biased and misleading results caused by the outliers, robust regression estimators have been presented as an alternative to the method of least squares. In this thesis study, the robust regression estimators which are less sensitive to outliers have been introduced by having been compared to the least squares estimators. In this context, with reference to the topics of the outlier detection methods, the stability criteria and the regression estimators, an application has been made in this field. Key Words: Least Squares, Outlier, Breakdown Point, Robust Regression, M Estimators.