PARAMETER ESTIMATION FOR LINEAR REGRESSION USING BOOTSTRAP METHOD

Authors : Recep Bindak

Pages : 1-5

View : 32 | Download : 20

Publication Date : 2018-12-30

Article Type : Research

Abstract :The bootstrap method firstly was introduced by Efron [1] as a general method for assessing the statistical accuracy of an estimator. Bootstrap is a computer-based re-sampling approach and a nonparametric statistical inference method. In this study, the use of the Bootstrap method in the parameter estimation of the linear regression is introduced and given a sample application on a real data set. In addition, if the data set contains outliers the effect that occurs in parameter estimation is examined. Confidence intervals and standard errors have been identified for various bootstrap repetitions numbers. As a result, it has been found that even 200 bootstrap repetations may suffice to obtain proper results.
Keywords : Bootstrap method, Resampling, Linear regression, Parameter estimation