Bootstrapping Residuals to Estimate the Standard Error of Simple Linear Regression Coefficients

Abstract

Regression models are the statistical methods that widely used in many fields. The models allow relatively simple analysis of complicated situations. The aim of the regression models is to analyze the relationship between the predictor and response. In order to do that, we have to estimate the regression coefficient. In case of simple linear regression, the method to estimate the regression coefficient is either least square method or maximum likelihood estimation. Also, the standard error of the regression coefficient is being estimated. In this paper, we apply the bootstrap method to estimate the standard error of the regression coefficient. We compare the result of the bootstrapping method with the least square method. From this study, we know that the standard error estimation value of regression model using the bootstrap method is close to the value if we use the least square method. So we can say that the bootstrap method can be used to estimate the standard error of another regression models coefficient which does not have the closed-form formula.

The application of Sugeno fuzzy to control active power load and remaining battery usage time modelling

“…This helps determine the hours taken for the maximum use of battery power from 55.82 Wh to 0 Wh. The basic linear regression (2) [25]- [27] used is:…”