Abstract

This article discusses the theoretical study and use excel and SPSS 19 application of Least Trimmed Square (LTS) methods and MM-estimation methods. Theoretical study focused on the elaboration of the concept of outlier, least trimmed square methods and MM-estimation methods and selection best model use the criteria R2 and resid value. Outlier is data on who did not attend a pattern common regression on the model produced, or not follow as a pattern data as a whole. The existence of outlier in the data can be disrupt the process of data analysis, that led to the data on residual and variance become larger. This research aims to know the effectiveness of robust regression method with Least Trimmed Square (LTS) and MM-estimation in multiple linear regression. This data consisting of age (X1) and body mass index (X2) as variable independent while systolik blood pressure (Y) as dependent variables. The model produced using Least Trimmed Square methods that is Y^=67.141+0.649X1+0.587X2. Regarding the resulting uses the method MM-estimation that is Y^=65.308+0.666X1+0.618X2. Because at Least Trimmed Square method (LTS) obtained the R2 value of is bigger and smaller than the residual method of MM-estimation then it can be concluded that the method of Least Square Trimmed (LTS) is more efficient in the estimate parameter of the regression compared the methods of MMestimation