COMPARISON OF ROBUST ESTIMATION ON MULTIPLE REGRESSION MODEL

This study aimed to compare the robustness of OLS method with a robust regression model on data that had outliers. The methods used were M-estimation, MM-estimation, and S-estimation. step taken was check characteristics against Furthermore, modeled without outliers using M-, MM-, S-estimations. results very different between outlier models in method. It reflected intercept standard error variables generated from models. Meanwhile, S-estimations quite stable able withstand presence Based three estimations outliers, MM-estimation best candidate because, addition having parameter estimation, it also smallest error, which 61.9 resulting model.