Herawati, Netti and Nisa, Khoirin (2010) ANALISISI REGRESI ROBUST MENGGUNAKAN METODE PENDUGA-MM. PROSIDING Seminar Nasional Sains dan Teknologi III. ISSN 978–979-8510-20-5

Nisa SATEK 2010.pdf

Download (3MB) | Preview


Ordinary least square (OLS) is a best linear unbiased estimator (BLUE) for regression analysis which can minimizes the mean square error. But it is very sensitive to outliers, even one extreme outlier can distort the prediction of OLS. MM-estimator is one of robust regression methods which is known as a high breakdown point and a high asymptotically efficiency estimator. In this paper we aim to look at the robustness of the regression coefficient of MM-Estimator. A simulation study was carried out using four sample sizes: n = 20, 60,100 and 200. We contaminated the data with 5%, 10%, 15%, 20%, 25% and 30% outliers. The effect of outliers on regression coefficient is studied by comparing the Mean Square Error (MSE) resulted by MM-estimator and by OLS in presence of outliers. The result shows that regression analysis using MM-Estimator is very robust.

Item Type: Article
Subjects: H Social Sciences > HA Statistics
Depositing User: DR. KHOIRIN NISA
Date Deposited: 27 Mar 2018 04:45
Last Modified: 27 Mar 2018 04:45
URI: http://repository.lppm.unila.ac.id/id/eprint/6620

Actions (login required)

View Item View Item