<![CDATA[The problems studied in this article is the selection of model formed from all possible subsets of independent random variables in multiple linear regression equation. The purpose of this study is to determine the multiple linear regression parameter estimation and choose the best regression equation which will be used to predict the random variables which are bound. In this Thesis, The method used is the the smallest Squares method and Bootstrap Methods. The smallest Squares method is used to determine the regression parameter estimation which the results of residual resampling used to obtain regression parameter estimation using the Bootstrap method. The Criteria of model selection uses Statistics . The equation regression with the smallest Statistics is the best equation model. Comparison between the estimation parameter and the parameter sought by Vector Norm. To be easy in calculating of determining parameter estimation and Statistics use Matlab program. The performance of the proposed method is viewed by simulation method. From the simulation result and comparison between the parameter estimation and the parameters, it can indicate that the Bootstrap method can estimate the parameter accurately. As the example of application, those both methods are applied in Indonesion Bank data relating to the amount of circulating money and the factors that influence it.]]>
Copyrights © 2011
