Embedded Files
Robust Weighted Least Squares Estimations Against Outliers and Heteroscedasticity
Robust Weighted Least Squares Estimations Against Outliers and Heteroscedasticity
I explore the performance of various robust regression methods under conditions of heteroscedasticity and outlier contamination. Using a synthetically generated dataset, I compare the Ordinary Least Squares (OLS) estimator with Weighted Least Squares (WLS), Iteratively Reweighted Least Squares (IRLS) using Huber and Tukey Bisquare M-estimators, and the Least Trimmed Squares (LTS) estimator.
Page updated
Google Sites
Report abuse