Dayan kl Regresyon Y ntemi ve e itli Sosyal Veriler üzerinde Ayk r G zlemlerin Te hisi ( Robust Regression Method and Diagnose Of Outliers on Several Social Data)

Robust Regression Method and Diagnose Of Outliers on Several Social DataFundamentals of the Research: In the presence of outlier(s), the method of Ordinary Least Squares (OLS) can be affected and gives misleading results. The common applications of OLS and the frequent existence of outliers in researches require to be tended to use resistant estimators in every statistical work.Purpose of the Research: Hence the purposes of this study are to introduce some high breakdown robust estimation techiques that are alternative to OLS, to investigate the effects of outliers on OLS and robust estimators and lastly diagnose of outliers.Main Discussion and Results: In case of having atypical and infrequent observations, OLS estimation results, standard errors and confidence intervals are affected badly. Whereas reliable results and outlier diagnosis can be attained with robust estimators. The consequences of robust and sensitive estimators on demographically structured data are analyzed within the context of this subject. These consequences are also ascertained in some of the data set which provide two-dimensional plots with a visual perception. Furthermore, diagnosed outliers in data verify the past demographic statistics' summaries.