부스트래핑(bootstrapping)기업을 활용한 회귀분석

초록

The purpose of this study is to present the bootstrap technique in the areas of public administration and policy as an alternative method to deal with statistical estimation problems in the conventional OLS regression analysis resulting from a limited number of observations. In particular, pairs bootstrapping that involves choosing random samples in paris from the original data set was applied. To assess and compare the accuracy and stability of the conventional OLS and bootstrap estimates, repeated random sampling that provides precise estimates of the coefficients was also conducted. The results of a Monte Carlo simulation showed that the bootstrap technique provided considerably accurate estimates of the parameters including R^2 and the adjusted R^2. One tentative conclusion can be drawn that the bootstrap technique could be used to solve the overfitting problem of R^2 in the OLS regression analysis.

키워드